1233104 (1) 玖、發明說明 【發明所屬之技術領域】 本發明係有關一種多値資料記錄及再生裝置，諸如光 碟裝置，其可被有效地利用以記錄及再生多値資料於_ _ 如光碟之記錄媒體上。 【先前技術】 習知的光學記錄及再生系統係一種光碟系統，其使用 雷射當作光源並以位元串執行記錄及再生，且包含一記錄 系統及一再生系統，其被視爲部分回應數位傳輸路徑。記 錄系統包含一調變編碼器，其將原始資料調變爲二元資料 、一記錄等化器，其產生二元中間資料以限制介於記錄系 統與再生系統之間的錯誤傳輸、及一半導體雷射驅動電路 。再生系統包含一讀取放大器、一波形等化器，其區分二 元中間資料自與記錄等化器所同步讀取之多値信號、一等 化器控制裝置，其控制波形等化器之分支（tap )係數以 自動地最佳化波形等化器之特性、及一解調編碼器，其再 生原始資料自區分的二元中間資料。日本公開專利申請案 編號-3 12〇1 8揭露上述光學記錄及再生系統，其記錄及再 生多値信號。 於此一光學記錄及再生系統中，記錄等化器可由一模 數（modulo )加法器所形成，而波形等化器可由一模數加 法器及一多値位準決定系統所形成，此多値位準決定系統 具有一依據等化器控制裝置之分支變數。波形等化器之多 -5- (2) 1233104 位準決定系統可由一具有變數分支系統之再生波形等化過 濾器、及一多値位準識別器所形成。 同時，資訊記錄方法已被建議以記錄多値資訊於一光 學資訊記錄媒體之資訊軌道上。於那些方法中，多値資訊 之記錄的執行係透過軌道方向上之資訊凹陷（pits )的長 度以及相對於軌道方向上之再生光點的資訊凹陷偏移（ shifting )量之各種組合。再者，符合上述方法之資訊再 生裝置已被提議。每一那些資訊再生裝置包含一多段光檢 測器、一儲存單元、及一資訊識別器。於此一資訊再生裝 置中，多段光檢測器檢測其反射自一光學資訊記錄媒體或 者穿透一光學資訊記錄媒體之再生光通量。儲存單元儲存 所有的光量及相應於多値資訊之資訊凹陷的光分佈，此多 値資訊係由軌道方向上之預定資訊凹陷的長度以及相關於 再生光點之預定資訊凹陷的位置所表示。資訊識別器透過 介於儲存單元中所儲存的光量與光分佈之間的相關（ correlations )以識別各資訊凹陷之資訊。日本公開專利 申請案編號5- 1 2 8 5 3 0揭露上述資訊再生裝置之型式的一 種範例。 然而，上述光學記錄及再生系統、上述資訊記錄方法 、及上述資訊再生裝置有數個問題。 上述光學記錄及再生系統利用一多値位準決定系統及 一模數相加器電路以當作一波形等化器。如圖1 7中所示 ，多値位準決定系統及模數相加器電路具有依據等化器控 制裝置之分支變數，以致其介於碼之間的干擾可被消除且 -6- (3) 1233104 資訊可被再生以高精確度。於此電路結構中，波形等化器 被執行爲輸入信號上之一線性操作（線性函數）。 圖1 8顯示多値記錄操作之範例，其中記錄標記之佔 有率係關於被稱爲”單元”之單位區域（areas )而被改變 。此操作將於下文中被稱爲”區域調變”。 以各記錄標記，則反射率變爲低於未記錄區域中之反 射率（亦即，”高至低記錄”）。如圖18中所示，區域（ 其中記錄標記不存在）中之一再生信號的信號輸出値（在 波形等化之前）被表示以編號①，0,及（3)之標記•。即使 單元中之記錄標記的佔有率相同，有於各相鄰記錄標記（ 如標示以①，Ο，及（3))的佔有率之間的差異而仍有信號輸 出値之間的差異。此係因爲介於記錄及再生點的直徑DM 與單兀長度（其目5錄及再生點以圖18中所不前頭之方向 執行掃瞄期間的時間週期）之間的關係爲DM > CL。因此 ，介於信號輸出値之間的差異可被視爲碼之間的干擾。 現在參考圖1 9 A至2 1 B，將描述碼間的干擾與各相鄰 記錄標記的佔有率之間的關係。 圖19A及19B顯示一種情況，其中記錄未被執行於 其位於一主體單元之前及之後的單元中（此情況將於下文 中被稱爲”孤立波（solitary wave)”情況）。圖20及20B 顯示一種情況，其中位於主體單元之前及之後的單元各具 有與主體單元之記錄標記佔有率相同的記錄標記佔有率（ 此情況將於下文中被稱爲”連續波”情況）。圖21A及 2 1 B顯示一種情況，其中位於主體單元之前及之後的單元 -7- (4) 1233104 均具有最高的記錄標記佔有率。 圖19A、20A及21A之圖形中的多値位準（〇至7) 指示記錄標記佔有率。更明確地，多値位準指示一單 元之未記錄狀態，而多値位準” 7”指示一具有最高記錄標 記佔有率之單元的狀態。於此，多値記錄係光學記錄。同 時，每一圖形具有•標記（其代表波形等化操作之前所測 得之値，以及具有一實線（其代表目標値）每一目標値係 一計算値，其代表其中波形干擾可被完全消除的情況，當 以圖1 7中所示之電路執行一波形等化操作時。 如圖1 9 A中所示，介於測量値與目標値之間的差異 展現相關於個別多値位準之一線性正比關係（一種線性關 係），且可藉由透過波形等化以改變線性操作之等化係數 (同等於圖I7中之常數C0至C4)而被校正。 至於圖20A，測量値係實質上與目標値相同，而因此 ，碼之間的干擾可透過波形等化而被消除。 另一方面，如圖2 1 A中所示，測量値與目標値之間 的差異於多位準〇至2之區域中是大的，且並未展現相對 於多値位準之線性正比關係。此證明其波形干擾無法由圖 1 7中所示之電路所完全消除。如上所述，於其中波形干 擾含有其不具線性影響之成分的情況下（例如，藉由參考 圖1 8所述之”區域調變”的多値記錄的情況），有一問題 ，亦即干擾無法藉由波形等化而被充分地消除。 爲了解決其干擾無法藉由上述資訊記錄方法及資訊再 生裝置而以波形等化來完全消除的問題，於一其中碼之間 -8 - (5) 1233104 的干擾含有其不具線性影響之成分的情況下（例如，藉由 參考圖1 8所述之”區域調變”的多値記錄的情況），則事 先透過所有組合型態以求取波形干擾之影響；對波形等化 操作加入校正；及利用其模擬人類資訊處理機制之神經網 路以當作一縮小求取程序中之各波形等化誤差的機構。 然而，需要一段可觀的期間以決定用來縮小各記錄標 記之誤差之收斂條件。結果，於其中藉由再生一***資料 區域中之求取區域以再生一未知資料區域的情況下，資料 再生速度無法被增加，雖然資料再生之可靠性可被增加。 【發明內容】 本發明之一般目的係提供一種多値資料記錄及再生裝 置及方法，其中係消除了上述缺點。 本發明之一更明確的目的係提供一種多値資料記錄及 再生裝置，其可透過一波形等化操作以準確地消除碼之間 的干擾’當再生其來自具有多値位準之再生信號的資訊時 。多値資料記錄及再生裝置調變一光學資訊記錄媒體上之 記錄標記的區域以再生多値資訊。 本發明之上述目的藉由一種多値資料記錄及再生裝置 而達成’此多値資料記錄及再生裝置依據一光學資訊記錄 媒體上之多値（0，1，2，…，（m - 1): m爲3或更大的 整數）資料以改變記錄標記之尺寸、及透過其藉由以一光 點掃瞄記錄標記而獲得之信號上的預定信號處理來檢測多 値資料。此多値資料記錄及再生裝置包含··一預測器，其 -9- (6) 1233104 預測性地決定多値資料；一延遲單元，其延遲預定信號處 理以其藉由預測器之預測性決定所需的時間週期；及一決 定器，其透過波形等化以決定多値資料，根據其爲來自預 測器之預測結果的預測性資料。 此多値資料記錄及再生裝置可進一步包含一波形等化 係數求取單元，其再生其中預錄有多値資料之光學資訊記 錄媒體上的一區域、決定此一預測性波形等化係數以致其 接受由一預測性波形等化電路所執行之信號處理的多値資 料之各項的信號輸出具有相關於目標値之最小可能誤差、 以及決定此一檢測性波形等化係數以致其接受由一檢測性 波形等化電路所執行之信號處理的多値資料之各項的信號 輸出具有相關於目標値之最小可能誤差。於此，檢測性波 形等化係數被決定於三個或更多連續記錄標記之各組合型 態，此等記錄標記包含一在主體記錄標記前之記錄標記串 的已知資料値、主體記錄標記之一已知資料値、及一接在 主體記錄標記後之記錄標記串的已知資料値。 此多値資料記錄及再生裝置可進一步包含一多値資料 檢測臨限値求取單元，其決定一用於預測多値資料之多値 資料檢測臨限値的預測性臨限値，根據其接受具有預測性 波形等化係數之波形等化的多値資料之各項的信號輸出、 及決定一用於最終地檢測多値資料之多値檢測臨限値的檢 測性臨限値，根據其接受具有檢測性波形等化係數之波形 等化的多値資料之各項的信號輸出。 此多値資料記錄及再生裝置可進一步包含一最終決定 -10- (7) 1233104 單元，其（當透過其根據預測性波形等化係數、檢測性波 形等化係數、預測性臨限値、及檢測性臨限値之信號處理 以再生一光學資訊記錄媒體上的未知多値記錄資料時）預 測性地決定多値資料，在以一僅用於預測之波形等化電路 執行波形等化之後、於等化條件下執行波形等化，這些條 件係依據其由檢測性波形等化電路參考三個或更多連續記 錄標記（包含一在主體記錄標記前之記錄標記串的預測値 、主體記錄標記之一預測値、及一接在主體記錄標記後之 記錄標記串的預測値）之組合型態所預測之組合型態而設 定、及最終地檢測來自再生信號之多値資料，其再生信號 係接受透過波形等化之信號處理。於此，預測値被包含於 其透過多値資料之預測性決定而獲得的預測資料中。 於此多値資料記錄及再生裝置中，多値資料之各項的 信號輸出可爲一信號輸出値，其可藉由再生三個或更多含 有與波形等化前相同之多値資料的連續記錄標記串而獲得 〇 於此多値資料記錄及再生裝置中，預測性波形等化電 路可爲一包含二個或更多分支之模數相加器電路。於此， 模數相加器電路係由一根據下列方程式而操作之電路所形 成： EQ(n) = D(n) + Σ { D (η) — D (η - i) } x C j i 其中係執行一波形等化操作於一來自第n記錄標記之 信號輸出上’ D(n)代表其執行於第^記錄標記上之波形等 -11 - (8) 1233104 化操作之前的信號輸出’ i及j隨分支之數目而改變，Cj 代表預測性波形等化係數’而EQ(n)代表在波形等化操作 之後的信號輸出。 於此多値資料記錄及再生裝置中，檢測性波形等化電 路可爲一包含三個或更多分支之模數相加器電路。於此， 模數相加器電路係由一根據下列方程式而操作之電路所形 成： EQ，（n) = D，（n) + Σ { D，（ η) — D，（ η - i)} x C j， i 其中係執行一波形等化操作於一來自第n記錄標記之 信號輸出上，D’（n)代表其執行於第n記錄標記上之波形 等化操作之後的信號輸出，i及j隨分支之數目而改變， Cj’代表檢測性波形等化係數，而EQ’（η)代表在波形等化 操作之後的信號輸出。 於此多値資料記錄及再生裝置中，預錄的已知多値資 料串可由一多値資料串所形成，其中包含多値資料之三個 或更多連續項的所有組合之資料串被重複地記錄；而光學 資訊記錄媒體可具有一記錄區域，其係分離自其中記錄有 未知多値資料之一資料區域。於此，記錄區域被週期性地 形成於光學資訊記錄媒體上，且記錄及再生被執行於記錄 區域中。 於此多値資料記錄及再生裝置中，預錄的已知多値資 料可由一資料串（其包含多値資料之三個或更多連續項的 所有組合）、以及一資料串（其中包含具有記錄資料之各 -12- (9) 1233104 項的信號輸出之相同目標値的多値資料之三個或更多連續 記錄標記串的型態被重複地記錄）所形成。於此，光學資 訊記錄媒體具有一記錄區域，其係分離自一其中記錄有多 値資料之資料區域。記錄區域被週期性地形成於光學資訊 記錄媒體上，且記錄及再生被執行於記錄區域中。 於此多値資料記錄及再生裝置中，檢測性波形等化係 數Cj’可被決定自下列方程式之5 (η):1233104 (1) Description of the invention [Technical field to which the invention belongs] The present invention relates to a multi-volume data recording and reproduction device, such as an optical disc device, which can be effectively used to record and reproduce multi-volume data on _ _ such as an optical disc On the recording medium. [Prior art] The conventional optical recording and reproduction system is an optical disc system that uses a laser as a light source and performs recording and reproduction with a bit string, and includes a recording system and a reproduction system, which are considered as a partial response Digital transmission path. The recording system includes a modulation encoder that converts the original data into binary data, a record equalizer, which generates binary intermediate data to limit erroneous transmission between the recording system and the reproduction system, and a semiconductor Laser drive circuit. The reproduction system includes a read amplifier, a waveform equalizer, which distinguishes the multiple intermediate signals read from the binary equalizer synchronously with the record equalizer, and a equalizer control device, which controls the branch of the waveform equalizer. (Tap) coefficients to automatically optimize the characteristics of the waveform equalizer and a demodulation encoder that reproduces the binary intermediate data from the self-distinguishing original data. Japanese Laid-Open Patent Application No.-31208 discloses the optical recording and reproducing system described above, which records and reproduces multiple signals. In this optical recording and reproduction system, the recording equalizer can be formed by a modulo adder, and the waveform equalizer can be formed by a modulo adder and a multi-level determination system. The level determination system has a branch variable based on the equalizer control device. The number of waveform equalizers -5- (2) The 1233104 level determination system can be formed by a regenerative waveform equalization filter with a variable branch system and a multi-level identifier. At the same time, information recording methods have been proposed to record multiple pieces of information on the information track of an optical information recording medium. In those methods, the execution of the multi-information recording is through various combinations of the length of the information pits in the track direction and the amount of information shifting relative to the reproduced light spot in the track direction. Furthermore, an information reproduction device conforming to the above method has been proposed. Each of those information reproduction devices includes a multi-segment light detector, a storage unit, and an information identifier. In this information reproduction device, the multi-segment photodetector detects the reproduction light flux reflected from an optical information recording medium or penetrating through an optical information recording medium. The storage unit stores all the light quantity and the light distribution of the information depression corresponding to the multi-information, which is represented by the length of the predetermined information depression in the track direction and the position of the predetermined information depression in relation to the reproduced light spot. The information identifier recognizes the information of each information depression through correlations between the amount of light stored in the storage unit and the light distribution. Japanese Laid-Open Patent Application No. 5- 1 2 8 5 3 0 discloses an example of the type of the above information reproduction device. However, the optical recording and reproducing system, the information recording method, and the information reproducing apparatus have several problems. The above-mentioned optical recording and reproducing system uses a multi-level determination system and an analog-to-digital adder circuit as a waveform equalizer. As shown in Figure 17, the multi-level determination system and the analog-to-digital adder circuit have branch variables based on the equalizer control device, so that the interference between codes can be eliminated and -6- (3 ) 1233104 Information can be reproduced with high accuracy. In this circuit structure, the waveform equalizer is performed as a linear operation (linear function) on the input signal. Fig. 18 shows an example of a multiple recording operation, in which the occupation rate of recording marks is changed with respect to a unit area (areas) called a "unit". This operation will be referred to as "area modulation" hereinafter. With each recording mark, the reflectance becomes lower than that in the unrecorded area (that is, "high to low recording"). As shown in FIG. 18, the signal output 値 (before the waveform is equalized) of one of the reproduced signals in the area (where the recording mark does not exist) is indicated with the marks ①, 0, and (3). Even if the occupancy rates of the recording marks in the unit are the same, there are differences between the occupancy rates of adjacent recording marks (such as ①, 0, and (3)), but there is still a difference between the signal output 値. This is because the relationship between the diameter DM of the recording and reproduction points and the unit length (the time period during which the 5 recording and reproduction points are scanned in a direction not shown in FIG. 18) is DM > CL . Therefore, the difference between the signal output 値 can be regarded as interference between codes. Referring now to FIGS. 19A to 21B, the relationship between the interference between codes and the occupancy rate of each adjacent recording mark will be described. 19A and 19B show a case in which recording is not performed in a unit that is located before and after a main unit (this case will be referred to as a "solitary wave" case hereinafter). Figures 20 and 20B show a case where the units before and after the main unit each have the same record mark occupancy rate as the record mark occupancy of the main unit (this case will be referred to as a "continuous wave" case hereinafter). Figures 21A and 21B show a case where the units before and after the main unit -7- (4) 1233104 have the highest record mark occupancy. The multiple levels (0 to 7) in the graphs of FIGS. 19A, 20A, and 21A indicate the recording mark occupancy. More specifically, the multi-level indicates the unrecorded status of a cell, and the multi-level "7" indicates the status of a cell having the highest record mark occupancy. Here, the multiple recording is an optical recording. At the same time, each figure has a • mark (which represents the 値 measured before the waveform equalization operation), and each target with a solid line (which represents the target 値) is a calculation 代表, which represents where the waveform interference can be completely eliminated. Eliminate the situation when a waveform equalization operation is performed with the circuit shown in Figure 17. As shown in Figure 19A, the difference between the measurement and target 値 manifestation is related to individual multi-levels A linear proportional relationship (a linear relationship) and can be corrected by changing the equalization coefficient (equivalent to the constants C0 to C4 in Figure I7) of the linear operation by equalizing the waveform. As for Figure 20A, the measurement system It is essentially the same as the target 値, and therefore, the interference between codes can be eliminated by waveform equalization. On the other hand, as shown in Figure 2 A, the difference between the measurement 値 and the target 値 is at multiple levels. The area from 0 to 2 is large and does not exhibit a linear proportional relationship with respect to the multi-level. This proves that its waveform interference cannot be completely eliminated by the circuit shown in Figure 17. As mentioned above, among them Waveform interference includes its non-linearity In the case of affected components (for example, the case of multi-field recording by "area modulation" described with reference to FIG. 18), there is a problem that the interference cannot be sufficiently eliminated by equalizing the waveform. Solve the problem that its interference cannot be completely eliminated by waveform equalization through the above information recording method and information reproduction device. In the case where the interference between codes -8-(5) 1233104 contains components that do not have a linear effect (For example, the case of multi-recording with "area modulation" as described with reference to FIG. 18), then through all combinations to obtain the effect of waveform interference in advance; add correction to the waveform equalization operation; and use The neural network, which simulates the human information processing mechanism, is used as a mechanism to reduce the equalization errors of the waveforms in the program. However, a considerable period is required to determine the convergence conditions for reducing the errors of each record mark. Results In the case where an unknown data area is reproduced by reproducing a seek area inserted into the data area, the data reproduction speed cannot be increased, although the data The reliability can be increased. [Summary of the Invention] The general purpose of the present invention is to provide a multi-data recording and reproducing device and method, in which the above disadvantages are eliminated. A more specific object of the present invention is to provide a multi-data Recording and reproducing device, which can accurately eliminate interference between codes through a waveform equalization operation, when reproducing information from a reproduction signal with multiple levels. The multiple data recording and reproduction device modulates an optical An area of a recording mark on an information recording medium is used to reproduce multiple pieces of information. The above object of the present invention is achieved by a multiple piece of data recording and reproducing device.値 (0, 1, 2, ..., (m-1): m is an integer of 3 or more) data to change the size of the recording mark, and a signal obtained by scanning the recording mark with a light spot Detect multiple signals on a predetermined signal process. This multi-channel data recording and reproduction device includes a predictor that predicts multi-channel data -9- (6) 1233104; a delay unit that delays a predetermined signal processing to predict its decision by the predictor The required time period; and a determinator, which determines multiple data through equalization of the waveform, according to which is predictive data of the prediction result from the predictor. The multi-channel data recording and reproduction device may further include a waveform equalization coefficient obtaining unit that reproduces a region on an optical information recording medium in which multi-channel data is pre-recorded, and determines the predictive waveform equalization coefficient such that The signal output of each item of multi-data that accepts the signal processing performed by a predictive waveform equalization circuit has the smallest possible error related to the target, and determines the equalization coefficient of this detective waveform so that it accepts a detection The signal output of the various data items of the signal processing performed by the waveform equalization circuit has the smallest possible error related to the target. Here, the detection waveform equalization coefficient is determined by each combination of three or more consecutive recording marks. The recording marks include a known data of the recording mark string before the main recording mark. The main recording mark. One of known data, and a known data of a record mark string following the main record mark. The multi-data recording and reproducing device may further include a multi-data detection threshold and obtaining unit, which determines a predictive threshold for predicting the number of multi-data and data detection thresholds, based on its acceptance. Signal output of each item of multi-data with equalization of predictive waveform equalization coefficients, and determination of a detection threshold for the final detection of multi-data, detection threshold, based on its acceptance Signal output of each item of multiple data with waveform equalization of detection waveform equalization coefficient. This multi-data recording and reproduction device may further include a final decision -10- (7) 1233104 unit, which (when the equalization coefficient based on the predictive waveform, the equalization coefficient of the detection waveform, the predictive threshold, and Detective threshold signal processing to reproduce an unknown multi-recorded data on an optical information recording medium) predictively determine multiple data, after performing waveform equalization with a waveform equalization circuit used only for prediction, Waveform equalization is performed under equalization conditions, based on which the detection waveform equalization circuit refers to three or more consecutive record marks (including a prediction of a record mark string before the main record mark), the main record mark One of the prediction type, and the prediction type of the recording mark string following the main recording mark, and the combination type predicted by the combination type is set, and finally multiple data from the reproduction signal are detected, and the reproduction signal is Signal processing through waveform equalization. Here, the prediction data is included in the prediction data obtained through the predictive decision of multiple data. In this multi-data recording and reproducing device, the signal output of each item of multi-data can be a signal output, which can reproduce three or more consecutive data files containing the same multi-data as before the waveform equalization. The mark string is obtained by recording. In this multi-data recording and reproducing device, the predictive waveform equalization circuit may be an analog-to-digital adder circuit including two or more branches. Here, the analog-to-digital adder circuit is formed by a circuit that operates according to the following equation: EQ (n) = D (n) + Σ {D (η) — D (η-i)} x C ji where A waveform equalization operation is performed on a signal output from the n-th record mark 'D (n) represents the waveform output performed on the ^ -th record mark, etc.-11-(8) 1233104 Signal output before the equalization operation' i And j vary with the number of branches, Cj represents the predictive waveform equalization coefficient 'and EQ (n) represents the signal output after the waveform equalization operation. In this multi-data recording and reproduction device, the detection waveform equalization circuit may be an analog-to-digital adder circuit including three or more branches. Here, the analog-to-digital adder circuit is formed by a circuit that operates according to the following equation: EQ, (n) = D, (n) + Σ {D, (η) — D, (η-i)} x C j, i where a waveform equalization operation is performed on a signal output from the n-th record mark, and D '(n) represents the signal output after the waveform equalization operation is performed on the n-th record mark, i And j vary with the number of branches, Cj ′ represents the detection waveform equalization coefficient, and EQ ′ (η) represents the signal output after the waveform equalization operation. In this multi-data recording and reproduction device, a pre-recorded known multi-data string can be formed from one multi-data string, in which a data string containing all combinations of three or more consecutive items of multi-data is repeatedly Recording; and the optical information recording medium may have a recording area which is separated from a data area in which unknown data is recorded. Here, the recording area is periodically formed on the optical information recording medium, and recording and reproduction are performed in the recording area. In this multi-data recording and reproduction device, the pre-recorded known multi-data can be composed of a data string (which contains all combinations of three or more consecutive items of multi-data), and a data string (which contains a record Each of the -12- (9) 1233104 items of signal output has the same target, and three or more types of continuous recording mark strings are repeatedly recorded). Here, the optical information recording medium has a recording area, which is separated from a data area in which multiple pieces of data are recorded. The recording area is periodically formed on the optical information recording medium, and recording and reproduction are performed in the recording area. In this multi-data recording and reproduction device, the detection waveform equalization coefficient Cj 'can be determined from 5 (η) of the following equation:
Cj’ = Ci- 5 (η) X Sj X {D(n) - D(n — i)} X G 其中5 (n)代表目標値與一再生信號之間的誤差，在 其執行於第η記錄標記的預測性波形等化操作之後；Cj 代表預測性波形等化係數；G代表一收斂增益；Sj係等於 c" Σ | Cj | ( I Cj I係Cj之絕對値）；且產生自此式 {D(n) — D(n - i)}之誤差係隨著Cj之等化係數的比例而改 〇 於此多値資料記錄及再生裝置中，收斂增益G可被 決定以使得其介於目標値與檢測性波形等化操作後的各信 號輸出之間的誤差被減至最小。 於此多値資料記錄及再生裝置中，預測性波形等化係 數之初始値可被預錄爲光學資訊記錄媒體上之系統資訊。 於此多値資料記錄及再生裝置中，預錄於光學資訊記 錄媒體上之預測性波形等化係數的初始値可被記錄爲多値 記錄資料，其具有較資料區域中所記錄之多値資料中的値 -13- (10) 1233104 數目更少的値數目。 於此多値資料記錄及再生裝置中，預錄於光學資訊記 錄媒體上之預測性等化係數的初始値以及其中預錄有已知 多値資料之區域可被再生，且預測性等化係數被接著決定 以使得其介於目標値與多値資料之各項的信號輸出之間的 誤差被最小化。 本發明之上述目的亦由一種多値資料記錄及再生裝置 達成，此種多値資料記錄及再生裝置依據一光學資訊記錄 媒體上之多値（0，1，2，…，（m-1) ··爲3或更大的整數 )資料而改變記錄標記之尺寸，並透過其藉由以一光點掃 瞄記錄標記所得之信號上的預定信號處理來檢測多値資料 。此多値資料記錄及再生裝置包含：一預測器，其預測性 地決定多値資料；一延遲單元，其延遲預定信號處理以其 藉由預測器之預測性決定所需的時間週期；及一決定器， 其透過波形等化以決定多値資料，根據其爲來自預測器之 預測結果的預測性資料。於此，延遲單元與決定器之組合 被重複地串聯配置，以使得其波形等化被重複直到多値資 料之決定結果收斂，以其被使用爲預測性資料而來自決定 器的決定結果。 此多値資料記錄及再生裝置可進一步包含一檢測性波 形等化係數求取單元，其再生其中預錄有已知多値資料之 光學資訊記錄媒體上的一區域、決定此一預測性波形等化 係數以致其接受由一預測性波形等化電路所執行之信號處 理的多値資料之各項的信號輸出具有相關於目標値之最小 -14 - (11) 1233104 可能誤差、以及決定此一檢測性波形等化係數以致其接受 由一檢測性波形等化電路所執行之信號處理的多値資料之 各項的信號輸出具有相關於目標値之最小可能誤差。於此 ，檢測性波形等化係數被決定於三個或更多連續記錄標記 之各組合型態，此等記錄標記包含一在主體記錄標記前之 記錄標記串的已知資料値及一接在主體記錄標記後之記錄 標記串的已知資料値，但排除待被再生之記錄標記的任何 已知資料値。 此多値資料記錄及再生裝置可進一步包含一多値資料 檢測臨限値求取單元，其決定一用於預測多値資料之多値 資料檢測臨限値的預測性臨限値，根據其接受具有預測性 波形等化係數之波形等化的多値資料之各項的信號輸出、 及決定一用於最終地檢測多値資料之多値檢測臨限値的檢 測性臨限値，根據其接受具有檢測性波形等化係數之波形 等化的多値資料之各項的信號輸出。於此，檢測性臨限値 之決定係根據其僅接受一次由檢測性波形等化電路之信號 處理的信號輸出値。 此多値資料記錄及再生裝置進一步包含一重複性處理 單元，其（當透過其根據預測性波形等化係數、檢測性波 形等化係數、預測性臨限値、及檢測性臨限値之信號處理 以再生一光學資訊記錄媒體上的未知多値記錄資料時）預 測性地決定多値資料，在以一僅用於預測之波形等化電路 執行波形等化操作之後、於等化條件下執行波形等化，這 些條件係依據其由檢測性波形等化電路參考三個或更多連 -15- (12) 1233104 續記錄標記（包含一在主體記錄標記前之記錄標記串的預 測値及一接在主體記錄標記後之記錄標記串的預測値，但 排除其透過預測性決定所獲得的預測性資料中待被再生之 主體記錄標記的任何已知資料値）之組合型態所預測之各 組合型態而被設定、使用其接受透過波形等化之信號處理 的再生信號以檢測多値資料、對多値資料執行多値決定、 及使用多値之決定結果當作預測性資料以重複波形等化直 到多値資料之決定結果收斂。 於此多値資料記錄及再生裝置中，多値資料之各項的 € 信號輸出之目標値可爲一信號輸出値，其可藉由再生三個 或更多含有與波形等化前相同之多値資料的連續記錄標記 串而獲得。 於此多値資料記錄及再生裝置中，預測性波形等化電 路可爲一包含三個或更多分支之模數相加器電路。於此， 模數相加器電路係由一根據下列方程式而操作之電路所形 成： EQ(n) = D(n) + Σ {D (η) - D (η - i)} x Cj i 其中係執行一波形等化操作於一第n記錄標記之信號 輸出上’ D(n)代表其執行於第^記錄標記上之波形等化操 作之前的信號輸出，i及j隨分支之數目而改變，Cj代表 預測性波形等化係數，而EQ(n)代表在波形等化操作之後 的信號輸出。 於此多値資料記錄及再生裝置中，檢測性波形等化電 -16- (13) 1233104 路可爲一包含三個或更多分支之模數相加器電路。於此， 模數相加器電路係由一根據下列方程式而操作之電路所形 成： EQ，（n) = D，（n)+ Σ {D5(n) - D5(n - i)} x Cj5 i 其中係執行一波形等化操作於第n記錄標記之信號輸 出上’ D’（n)代表其執行於第n記錄標記上之初始波形等 化操作之後的信號輸出，i及j隨分支之數目而改變，Cj， 代表檢測性波形等化係數，而EQ’（η)代表在一後波形等 化操作之後的信號輸出。 於此多値資料記錄及再生裝置中，預錄的已知多値資 料串可由一多値資料串所形成，其中包含多値資料之三個 或更多連續項的所有組合之資料串被重複地記錄；而光學 資訊記錄媒體可具有一記錄區域，其係分離自其中記錄有 未知多値資料之一資料區域。於此，記錄區域被週期性地 形成於光學資訊記錄媒體上，且記錄及再生被執行於記錄 區域中。當待被再生多値資料被決定時，求取資訊係透過 一統計操作而被更新，其中最近的求取結果被加至已被記 錄及再生之求取結果。 於此多値資料記錄及再生裝置中，預錄的已知多値資 料可由一資料串（其包含多値資料之三個或更多連續項的 所有組合）、以及一資料串（其中包含具有記錄資料之各 項的信號輸出之相同目標値的多値資料之三個或更多連續 記錄標記串的型態被重複地記錄）所形成。於此，光學資 -17- (14) 1233104 訊記錄媒體具有一記錄區域，其係分離自一其中記錄有多 値資料之資料區域。記錄區域被週期性地形成於光學資訊 記錄媒體上，且記錄及再生被執行於記錄區域中。當待被 再生多値資料被決定時，求取資訊係透過一統計操作而被 更新，其中最近的求取結果被加至已被記錄及再生之求取 結果。 於此多値資料記錄及再生裝置中，檢測性波形等化係 數Cj5可被決定自下列方程式之5 (η):Cj '= Ci- 5 (η) X Sj X {D (n)-D (n — i)} XG where 5 (n) represents the error between the target 値 and a reproduced signal, and it is executed at the n-th record After the labeled predictive waveform equalization operation; Cj represents the predictive waveform equalization coefficient; G represents a convergence gain; Sj is equal to c " Σ | Cj | (I Cj I is the absolute 値 of Cj); and is generated from this formula The error of {D (n) — D (n-i)} is changed with the proportion of the equalization coefficient of Cj. In this multi-data recording and reproduction device, the convergence gain G can be determined so that it is between The error between the target chirp and each signal output after the detection waveform equalization operation is minimized. In this multi-data recording and reproduction device, the initial value of the predictive waveform and other coefficients can be pre-recorded as system information on the optical information recording medium. In this multi-data recording and reproduction device, the initial value of the predictive waveform equalization coefficient pre-recorded on the optical information recording medium can be recorded as multi-data recording data, which has more data than the data recorded in the data area. The 値 -13- (10) 1233104 number is smaller in the number of 値. In this multi-data recording and reproduction device, the initial value of the predictive equalization coefficient pre-recorded on the optical information recording medium and the area in which the known multi-data is pre-recorded can be reproduced, and the predictive equalization coefficient is It is then decided to minimize the error between the signal output of each item of the target data and the multiple data. The above-mentioned object of the present invention is also achieved by a multi-channel data recording and reproducing device, which is based on the multi-channel (0, 1, 2, ..., (m-1) on an optical information recording medium). ···················· The size of the recording mark is changed for data of an integer of 3 or more, and multiple pieces of data are detected by a predetermined signal processing on the signal obtained by scanning the recording mark with a light spot. The multiple data recording and reproduction device includes: a predictor that predictively determines multiple data; a delay unit that delays a predetermined signal processing to a time period required by the predictive decision by the predictor; and The determinator, which determines multiple data by equalizing the waveform, is based on the predictive data of the prediction result from the predictor. Here, the combination of the delay unit and the determinator is repeatedly arranged in series so that its waveform equalization is repeated until the decision result of multiple data converges, and it is used as predictive data from the decision result of the determinator. The multi-frame data recording and reproduction device may further include a detection waveform equalization coefficient obtaining unit that reproduces a region on an optical information recording medium in which known multi-frame data is pre-recorded, and determines the prediction waveform equalization. Coefficient so that it accepts signal processing by a predictive waveform equalization circuit. The signal output of each item of data has a minimum value related to the target. -14-(11) 1233104 Possible errors and determine this detectability The waveform equalization coefficient is such that the signal output of each item of the multi-data that accepts the signal processing performed by a detection waveform equalization circuit has the smallest possible error related to the target. Here, the detection waveform equalization coefficient is determined by each combination of three or more consecutive recording marks. The recording marks include a known data of a recording mark string before the main recording mark, and a continuous The body records the known data of the record mark string after the mark, but excludes any known data of the record mark to be reproduced. The multi-data recording and reproducing device may further include a multi-data detection threshold and obtaining unit, which determines a predictive threshold for predicting the number of multi-data and data detection thresholds, based on its acceptance. Signal output of each item of multi-data with equalization of predictive waveform equalization coefficients, and determination of a detection threshold for the final detection of multi-data, detection threshold, based on its acceptance Signal output of each item of multiple data with waveform equalization of detection waveform equalization coefficient. Here, the detection threshold 临 is determined based on the signal output 接受 which accepts the signal processing by the detection waveform equalization circuit only once. This multi-data recording and reproduction device further includes a repetitive processing unit which (when passing the equalization coefficient based on the predictive waveform, the equalization coefficient of the detective waveform, the predictive threshold, and the detectable threshold signal When processing to reproduce an unknown multi-recorded data on an optical information recording medium) predictively determine the multi-view data, and execute the waveform equalization operation with a waveform equalization circuit for prediction only, and execute under equalization conditions Waveform equalization, these conditions are based on the detection of the waveform equalization circuit by referring to three or more consecutive -15- (12) 1233104 continuation record marks (including a prediction of a record mark string before the main record mark) and a Prediction of the record mark string following the main record mark, but excluding any known data of the main record mark to be reproduced from the predictive data obtained through the predictive decision, i.e. each predicted by the combination type Combined types are set, using the reproduced signal that receives signal processing through waveform equalization to detect multiple data, perform multiple decisions on multiple data, And use the decision result of multiple data as predictive data and repeat the waveform equalization until the decision result of multiple data converges. In this multi-data recording and reproduction device, the target of signal output of each item of multi-data can be a signal output, which can be reproduced by three or more containing the same as before waveform equalization.値 Obtained by continuously recording tag strings of data. In this multi-data recording and reproducing device, the predictive waveform equalization circuit may be an analog-to-digital adder circuit including three or more branches. Here, the analog-to-digital adder circuit is formed by a circuit that operates according to the following equation: EQ (n) = D (n) + Σ {D (η)-D (η-i)} x Cj i where A waveform equalization operation is performed on the signal output of the nth record mark. D (n) represents the signal output before the waveform equalization operation performed on the ^ th record mark. I and j change with the number of branches. , Cj represents the predictive waveform equalization coefficient, and EQ (n) represents the signal output after the waveform equalization operation. In this multi-data recording and reproduction device, the detection waveform equalization circuit -16- (13) 1233104 can be an analog-to-digital adder circuit including three or more branches. Here, the analog-to-digital adder circuit is formed by a circuit that operates according to the following equation: EQ, (n) = D, (n) + Σ {D5 (n)-D5 (n-i)} x Cj5 i where a waveform equalization operation is performed on the signal output of the nth record mark 'D' (n) represents the signal output after the initial waveform equalization operation performed on the nth record mark, i and j follow the branch The number changes, Cj, represents the detection waveform equalization coefficient, and EQ '(η) represents the signal output after a waveform equalization operation. In this multi-data recording and reproduction device, a pre-recorded known multi-data string can be formed from one multi-data string, in which a data string containing all combinations of three or more consecutive items of multi-data is repeatedly Recording; and the optical information recording medium may have a recording area which is separated from a data area in which unknown data is recorded. Here, the recording area is periodically formed on the optical information recording medium, and recording and reproduction are performed in the recording area. When multiple pieces of data to be reproduced are determined, the seeking information is updated through a statistical operation, where the most recent seeking result is added to the seeking result that has been recorded and reproduced. In this multi-data recording and reproduction device, the pre-recorded known multi-data can be composed of a data string (which contains all combinations of three or more consecutive items of multi-data), and a data string (which contains a record The signal output of each item of the data is the same as that of three or more types of continuous recording mark strings of multiple data, which are repeatedly recorded). Here, the optical recording medium has a recording area, which is separated from a data area in which multiple data are recorded. The recording area is periodically formed on the optical information recording medium, and recording and reproduction are performed in the recording area. When multiple pieces of data to be reproduced are determined, the seeking information is updated through a statistical operation, where the most recent seeking result is added to the seeking result that has been recorded and reproduced. In this multi-data recording and reproduction device, the detection waveform equalization coefficient Cj5 can be determined from 5 (η) of the following equation:
Cj，= Ci - 5 (η) X Sj X {D(n) - D(n - i)} X G 其中5 (n)代表目標値與一再生信號之間的誤差，在 其執行於第η記錄標記的預測性波形等化操作之後；Cj 代表預測性波形等化係數；G代表一收斂增益；Sj係等於 Cj/ Σ | Cj | ( I Cj I係Cj之絕對値）；且產生自此式 {D(n) - D(n - i)}之誤差係隨著Cj之等化係數的比例而改 變 〇 於此多値資料記錄及再生裝置中，收斂增益G可被 決定以使得其介於目標値與檢測性波形等化操作後的各信 號輸出之間的誤差被減至最小。 於此多値資料記錄及再生裝置中，預測性波形等化係 數之初始値可被預錄爲光學資訊記錄媒體上之系統資訊。 於此多値資料記錄及再生裝置中，預錄於光學資訊記 錄媒體上之預測性波形等化係數的初始値可被記錄爲多値 -18- (15) 1233104 記錄資料，其具有較資料區域中所記錄之多値資料中的値 數目更少的値數目。 於此多値資料記錄及再生裝置中，預錄於光學資訊記 錄媒體上之預測性等化係數的初始値以及其中預錄有已知 多値資料之區域可被再生，且預測性等化係數被接著決定 以使得其介於目標値與多値資料之各項的信號輸出之間的 誤差被最小化。 本發明之上述及其他目的、特徵及優點將由下列詳細 說明配合上後附圖形而變得更淸楚明白。 【實施方式】 以下參考後附圖形以描述本發明之實施例。 如先前技術中所述，一習知的波形等化操作無法通過 波形等化以充分地消除波形干擾，於其中波形干擾含有其 不具線性影響之成分的情況下（藉由圖1 8所述之”區域 調變”的多値記錄）。如參考圖19A至21B所述，碼間的 干擾展現不同程度的影響，根據位於待被再生目標之主體 記錄標記前及後的記錄標記之狀態。 爲了解決此問題，檢測性波形等化操作係依據位於主 體記錄標記（其爲一待被再生目標）前及後之記錄標記的 狀態以及影響之程度而被重複。欲如此做，則必須預測多 値資料。然而，因爲其來自波形等化前之一再生信號爲多 少模糊的（indistinct )，故必須執行一操作以增加資料 預測之準確性。當作用以增加資料預測之準確性的操作， -19- (16) 1233104 一使用固定等化常數之波形等化操作被執行於再生信號上 ，以使得其多値資料可被準確地預測自再生信號。 爲了依據依據位於主體記錄標記（其爲一待被再生目 標）前及後之記錄標記的狀態及影響之程度以執行一波形 等化操作’則必須設定每一三個或更多組合型態，包含主 體記錄標記之前的記錄標記串、主體記錄標記、及主體記 錄標記之後的記錄標記串。 例如，當一記錄及再生點係位於主體記錄標記之中心 時，以其介於記錄及再生點的直徑BD與單元長度CL之 間的關係約略爲BD = 2 X CL，則記錄及再生點便跨越三 個記錄標記。因此，必須參考其用於來自預測多値資料之 每一三個組合型態中的波形等化之等化係數以使波形等化 條件最佳化，並於最佳化之波形等化條件下執行檢測性波 形等化。 上述波形等化條件可隨著資訊記錄裝置與光學資訊記 錄媒體之組合而改變，根據光學資訊記錄裝置間之記錄條 件的變化以及記錄媒體間之記錄辨別力（sensibinty)的 改變。 因此，爲了有效地依據本發明以增加波形等化性能， 則必須事先再生已知的多値資料、求取預測性波形等化條 件及檢測性波形等化條件、並接著根據求取結果以再生未 知的多値資料。 圖1顯示一種多値資料檢測電路之結構，其係依據本 發明之多値資料記錄及再生裝置的一實施例。 -20- (17) 1233104 此多値資料檢測電路包含一預測性波形等化器11、 一預測器1，其從一接受預測性波形等化之再生信號預測 多値資料、一檢測性波形等化器3 1、及一決定器3，其從 一接受檢測性波形等化之再生信號預測多値資料。 以此電路結構，則波形等化之條件被求取，且未知的 資料被再生。 於此電路結構中，一初始等化係數被事先記錄於一光 學資訊記錄媒體上以獲得預測性波形等化之較高的求取效 率。藉由此動作，則可消除其由於光學資訊記錄媒體間之 等化係數的改變所需之長求取時間的問題。 再者，爲了增加預測結果之準確性，波形等化之一收 斂目標値被產生自一多値記錄的再生信號。因此，即使再 生信號隨著諸如反射率改變或記錄辨別力改變等改變而波 動，仍得以預測具有再生信號波動之多値資料。 爲了求取目標値，使用下列型態是有效的，考量其波 形等化之特性。 圖2及3係說明其中執行八進記錄波形等化之情況中 的收斂狀態之圖形。 介於記錄及再生點直徑BD與單元長度CL之間的關 係被約略表示爲：BD = 2 X CL。如圖2及3中所見，多 値資料之各項的收斂値係包括（0 0 0 )，（ 1 1 1 )，（ 2 2 2 )，（ 3 3 3 ) ，（444)，（5 5 5 )，（666)，及（777)之（\，丫，2)的組合型態 ，其中X代表主體記錄標記之前的多値資料、y代表主體 記錄標記之多値資料、而z代表主體記錄標記之後的多値 -21 - (18) 1233104 資料。 根據上述原理，目標値可簡單地藉由再生其以八種型 態（0 0 0)至（7 7 7)之組合形式所記錄之多値資料的型態而被 決定。 於八進記錄之情況下，（X，y，Z)之組合型態的總數 爲83 = 5 1 2。因此，目標値可被求取自僅有八種型態之組 合，其爲極有效率的。於此，“ 0”代表其中未記錄有記錄 標記之單元，而“7”代表其中記錄有最多記錄標記之單元 〇 爲了有效地收斂多値資料至八個目標値，則使用圖4 中所示之波形等化電路以取代圖1 7中所示之習知技術的 波形等化電路。 圖4中所示之波形等化電路具有一種5分支的結構。 此結構被設計以消除波形干擾自主體記錄標記前之兩個記 錄標記及主體記錄標記後之兩個記錄標記，其總共有四個 記錄標記。 如上所述，當記錄及再生點被置於主體記錄標記之中 心且跨越三個記錄標記時，則主體記錄標記接收碼之間的 千擾，其爲來自主體記錄標記之前及之後的兩個記錄標記 之影響。然而，爲了增加以波形等化器之校正準確性，更 理想的是一種5分支的結構，其可減少來自主體記錄標記 之前及之後的記錄標記之干擾。 於圖4所示之結構中，四個差異計算器42a至42d之 所有輸出爲，，〇 ”，當五個連續單元具有相同的多値資料時 -22- (19) 1233104 因此，當目標値（x，y，z)之組合爲x = y = z時，其爲 用以求取目標値之型態，則四個差異計算器42 a至42 d爲 ” 〇”，而不管波形等化係數之値。因此，圖4中所示之結 構作用爲一收斂於目標値上之波形等化器。 同時，此波形等化電路具有四個等化係數，其較圖 1 7中所示之習知電路少了 一個等化係數。因此，此結構 更有利在於求取等化係數時有更少的計算步驟。 接下來，將描述藉由一種利用圖1之多値資料檢測電 路的資訊記錄及再生電路之一多値資料信號的記錄及再生 〇 圖5係顯示一種利用圖1所示之多値資料檢測電路的 資訊記錄及再生電路之結構的一方塊圖。此資訊記錄及再 生電路執行一光學資訊記錄媒體50上之資料記錄及再生 ，並包含一拾取頭51、一 LD驅動器信號產生器52、一 多値資料轉變器5 3、一資訊資料產生器5 4、一光檢測器 55、一 AGC控制器56、一同步信號單元檢測電路57、一 取樣信號產生電路58、定量（quantum) AD轉變器59、 及一多値信號記憶體60。 資訊資料產生器54產生光學資訊記錄媒體50上之待 被記錄數位資料，而多値資料轉變器5 3將多値資料轉變 爲八進値。 假如，舉例而言，數位資料爲” 〇 〇 1 1 〇 1 〇 1 〇，，，則多値 資料轉變器5 3將數位資料之各三個圖形轉變爲一八進數 -23- (20) 1233104 字，並產生八進資料” 1 52”。根據八進資料，一待被記錄 以LD驅動器信號產生器52之記錄脈衝型態被產生，且 拾取頭51之半導體雷射光源（LD)被驅動以致其拾取頭 51將雷射光束L聚集於旋轉中的光學資訊記錄媒體5 〇上 。因此，光學記錄被執行於光學資訊記錄媒體5 〇上。 光學資訊記錄媒體5 0可爲一種於其上使用諸如”〜 次寫入”材料之顏料材料的CD-R碟片，或爲一種於其上 使用可重寫相位可變材料的CD-RW碟片。 圖6顯示圖1所示之光學資訊記錄媒體5 〇的記錄資 料格式。 如圖6中所示，光學資訊記錄媒體50之記錄資料格 式包含被稱爲”區段（sectors )，，之單位區塊，諸如”求 取區段”及”資料區段1至N”。每一區段包含一區段標記 、一同步化信號區域、一位址區域、及一多値資料區域。 區段標記指示區段之開始，且係由未出現於資料區域 中之一型態（未顯示）所形成，諸如” 00000000777777，， 〇 同步化信號區域係由一重複型態（未顯示）所形成， 諸如” 〇7〇7〇7〇7〇7〇7〇7 07...。重複型態將被使用爲一取樣 時鐘，當多値資料被量化時。位址區域指示區段之位址， 而位址被記錄爲多値資料。 爲了再生一信號，雷射光被聚集於旋轉中之光學資訊 記錄媒體50上，且反射回的光藉由光檢測器5 5而被光電 地轉變爲電信號。AGC控制器5 6檢測來自再生信號之一 (21) 1233104 區段標記，並進一步檢測區段標記之最大信號値” 00000000”及最小信號値” 77777777”。AGC控制器56接 著執行自動增益控制（AGC )以穩定振幅（亦即，介於最 大信號値與最小信號値之間的差異）。AGC被執行以校 正由於光學資訊記錄媒體5 0造成之反射率變異所導致的 不正確多値決定。 同步信號單元檢測電路5 7接下來檢測一同步化信號 區域。根據同步化信號，則取樣信號產生電路5 8產生一 計時信號。定量AD轉變器59接著以計時信號量化（類 比至數位轉變）多値信號（以樣本固持再生信號於單元之 中心）。量化的多値信號被連續地記錄於多値信號記憶體 6〇中。多値信號記憶體6〇中所記錄之信號被連續地讀取 並輸入爲量化的再生信號於圖1所示之多値資料檢測電路 中。信號將被用於波形等化及多値決定。 接下來，將描述其求取多値資料檢測電路中之預測性 等化係數的程序。 圖1之初始步驟單元（亦即，預測性波形等化器11 及多級資料預測器1 6 )被使用爲預測性波形等化之一電 路結構。同時，於此利用一種具有如圖1中所示之電路結 構的波形等化器。 圖4中所示之波形等化器根據下列方程式（7)以執行 一計算操作： •25- (22) 1233104 EQ(n) = D(n)+{D⑻一 D(n - 2)}x C0+{D⑻-(η - l)}x C1+{D⑻-(η + 1)}χ C2+{D⑻-(n + 2)}xC3....(7) 於此方程式中，D(n)代表第 η單元之再生信號， EQ(n)代表在預測性波形等化操作之後的信號輸出値，而 C0至C3代表等化係數。這些等化係數爲固定的係數，無 論其多値資料型態。 圖7係求取預測性等化係數之操作演算法的流程圖。 一已知的資料信號具有一求取區段以用於每一 Ν連 續的資料區段，如圖6之格式中所示。求取區段中之多値 資料爲已知的資料。 已知資料係由其中所有組合型態（5 1 2 - 8 = 5 04型態 )被重複地記錄如目標値型態之資料所形成。初始等化係 數被預錄於一分離自資料記錄區域之區域中，而資訊將藉 由再生分離的區域而被獲得。 然而’假如波形等化操作並未最佳化，則無法檢測到 準確的多値資料。因此，更理想的是使其初始等化係數被 記錄爲二元資訊，以致其初始化係數可被準確地獲得。 預測性等化係數之決定被重複直到關於目標信號之誤 差變爲等於一預定値（Τ0 )或更小。欲計算誤差，則必須 獲得有關已知資料之多値資料的配置的資訊。有關已知資 料之配置的資訊被記錄於圖5所示之資訊記錄及再生裝置 中的記憶體（未顯示）中，且被使用以記錄及再生求取資 -26- (23) 1233104 料。圖5中所示之多値信號記憶體60需一資料容量等於 或大於其相應於用以決定等化係數所需之時間週期的尺寸 。具有足夠大的資料容量，則將於預測性等化係數決定操 作期間及後續檢測性等化係數決定操作期間被再生之未知 的多値資料可被暫時地儲存，而資料處理可被連續地執行 。在預測性等化係數決定之後，於接收到已完成預測性等 化係數決定之通知時開始求取檢測性等化係數。 於已知資料之再生時，求取結果（亦即，決定的預測 性等化係數CO，Cl，C2,及C3)被設定至圖1中所示之 預測性等化係數計算器1 2，並接著接受波形等化。檢測 性波形等化中所利用之預測性資料爲由多値資料預測器 1 6所獲得的多値資料，此多値資料預測器i 6係轉變其接 受預測性波形等化之信號。於此，決定臨限値被決定自檢 測的目標値資料。目標値被記錄於一預測性臨限値產生器 1 7中，且被多値資料預測器1 6使用爲臨限値資訊。 求取結果（亦即，預測性等化係數C 0，C 1，C 2，及 C 3，以及根據目標値之多値資料決定臨限値）被更新於每 次圖6中所示的求取區段之一被再生時，且被使用以再生 未知的多値資料。 接下來，將描述於多値資料檢測電路中求取檢測性等 化係數之程序。 圖1之後半單元（亦即，檢測性波形等化器3 1及多 値資料檢測器3 4 )被使用爲一用於檢測性波形等化之電 路結構。同時，一種具有如圖4中所示之電路結構的波形 -27- (24) 1233104 等化器被利用於此。 圖4中所示之波形等化器根據下列方程式（8)以執行 一計算操作： EQ，(n) = D(n)+{D(n) - D(n - 2)}x CO (I，J，K) +{D(n) - (η - l)}x Cl (I 小 K) +{D(n) - (η + l)}x C2 (I，J，K) +{D⑻-(η + 2)}x C3 (I，J，K).·..(8) 於此方程式中，D(n)代表第 η單元之再生信號， EQ’（η)代表在檢測性波形等化操作之後的信號輸出値。於 此，而CO (I ’ J，Κ)至C3 (I，J，Κ)代表相應於各S!]型態 (I，J，Κ)之等化係數。此外，I代表多値資料之第（n-i) 項，：f代表多値資料之第η項，而K代表多値資料之第 (η + 1 )項。 圖8係求取檢測性等化係數之操作演算法的流程圖。 每一已知資料信號具有如圖6中所示之格式，其係相 同於預測性波形等化。已知資料係由其中所有組合型態（ 5 1 2 - 8 = 5 04型態）被重複地記錄如目標値型態之資料所 形成。 爲了再生已知資料並計算波形等化係數以致其 E Q ’（ η)之値收斂於目標値，則至少四個連續單元之結果（ 例如，EQ(n-2)，EQ(n-l)，EQ(n)，及 EQ(n+l)之結果）需 根據方程式（7)而被計算。因此，必須求出同步方程式並 獲得每一 5 1 2型態（xyz )之最佳波形等化係數。此涉及 大量的計算，且需要極長的求取時間。 -28- (25) 1233104 然而，關於預測性波形等化時所產生之目標値的誤差 係直接正比於其根據方程式（7)之預測性波形等化所決定 的等化係數CO，Cl，C2,及C3。 因此，當多値資料之第η項被再生時之相應目標値的 相關誤差爲5 (η)時，則可建立下列方程式（9)至（12)。 從{D(n)-D(n-2)}產生之誤差〇c C0/ {|C0| + |Cl| + |C2| + |C3|}x a(n) …（9 ) 從{0(11)-0(11-1)}產生之誤差〇〇(：1/ {|C0| + |Cl| + |C2| + |C3|}x a(n) ."(10) 從⑺㈠丨^⑶+:^丨產生之誤差^匸二/ {|C0| + |Cl| + |C2| + |C3|}x a(n) …（1 1 ) 從{D(n)-D(n + 2)}產生之誤差〇c C3/ {|C0| + |Cl| + |C2| + |C3|}x a(n) - ( 12 ) 從以上的方程式（9)至（l2)，則相關於當多値資料之第 η項被再生時之相應目標値的誤差3 (η)被決定。此外，檢 測性波形等化中之最佳等化係數C0 (I，J，Κ)、C1 (I，J ，Κ)、C2 (I，J，Κ)、及C3 (I，J，Κ)可藉由下列方程式 (1 3 )至（1 6)而被計算，其係使用誤差5 (η)及預測性等化係 數CO，Cl，C2,及C3。應注意I代表多値資料之第（η-1} 項’ J代表多値資料之第η項，而Κ代表多値資料之第 (η + 1)項。 -29 - (26) 1233104 C0(I，J，K) = C0-a(n)x SOx {D(n)-D(n-2)}x G …（13) C1(I，J，K) = Cl-a(n)x Six {D(n)-D(n-l)}x G …（14) C2(I，J，K) = C2-a(n)x S2x {D(n)-D(n+l)}x G ··· (15) C3(I，J，K) = C3-a(n)x S3x {D(n)-D(n + 2)}x G ··· (16) 於以上方程式中，G代表收斂增益。同時，SO，SI， S2,及S3係由下列方程式（17)至20)所界定，其中I P I代 表P之絕對値。 so =co/{ l CO | - H Cl 1 + | C2 1 + 1 C3 1 }. ..(17) S 1 =Cl/{ l CO | - M ci | + | C2 1 + 1 C3 1 • (18) S2 =C2/{ 1 CO 1 1 M ci | + | C2 1 + 1 C3 1 • (19) S3 =C3/{ | CO 1 H H ci | + | C2 1 + 1 C3 1 }. ..(20) 値 C0 (I，J，K)、C1 (I，J，K)、C2 (I，J，K)、及 C3 (I，J，K)可透過根據方程式（13)至（16)之操作而被決 定自6 (η)。因此，無需求出同步方程式。 以此方式，最佳等化係數可被自動地決定自多値資料 之第η項的誤差5 (η)以及預測性等化係數CO，Cl，C2, 及C 3。因此，檢測性等化係數之計算時間可被顯著地縮 檢測性等化係數之決定被重複直到相對於相應目標値 之誤差變爲等於一預定値（Τ0’）或更小。 於未知資料之再生時，求取（亦即，已決定的檢測性 -30- (27) 1233104 等化係數 C0 (I，J，K)、Cl (I，J，K)、C2 (I，J，κ)、 及C3 (l，j，K)、以及收斂增益G)被設定至圖1中之檢 測性等化係數及收斂增益計算器32，並接著執行波形等 化。 圖9顯示用於檢測波形等化之等化係數的一表列。 檢測性等化係數被設定並儲存於檢測性等化係數及收 被增益g十算器3 2中，根據再生的已知資料。如圖9中所 示’檢測性等化係數之設定被執行於三個連續型態之每一 單元。 依據其藉由多値資料預測器1 6所獲得的多値資料， 最佳的等化係數及收斂增益被讀取自檢測性等化係數及收 斂增益計算器3 2。之後，檢測性波形等化被執行，接著 由多値資料檢測器3 4執行多値決定。爲了以高精確度決 定多値資料，更理想的是使用檢測性波形等化時所計算的 多値信號之求取結果，其優於使用從預測性臨限値所求取 之目標値資料。 因此，當已知資料被再生及求取時，各多値位準上之 信號分佈接受其根據多値資料及再生信號之結果的統計處 理，而所得的信號被使用爲下次多値決定之臨限値。 因此，上述臨限値之統計計算操作被執行以圖1中所 示之多値信號記憶體3 6及統計處理器3 7，而所得的臨限 値被設定至檢測性臨限値產生器3 5。使用這些臨限値， 則檢測性波形等化之後的多値信號可被準確地決定。求取 結果（亦即，已決定的檢測性等化係數C0 (I，J，K)、C 1 -31 - (28) 1233104Cj, = Ci-5 (η) X Sj X {D (n)-D (n-i)} XG where 5 (n) represents the error between the target 値 and a reproduced signal, and it is executed at the η record After the labeled predictive waveform equalization operation; Cj represents the predictive waveform equalization coefficient; G represents a convergence gain; Sj is equal to Cj / Σ | Cj | (I Cj I is the absolute 値 of Cj); and is generated from this formula The error of {D (n)-D (n-i)} changes with the ratio of the equalization coefficient of Cj. In this multi-data recording and reproduction device, the convergence gain G can be determined so that it is between The error between the target chirp and each signal output after the detection waveform equalization operation is minimized. In this multi-data recording and reproduction device, the initial value of the predictive waveform and other coefficients can be pre-recorded as system information on the optical information recording medium. In this multi-volume data recording and reproduction device, the initial value of the predictive waveform equalization coefficient pre-recorded on the optical information recording medium can be recorded as multi-volume -18- (15) 1233104 recording data, which has a larger data area The larger number of records in the record, the smaller the number of records. In this multi-data recording and reproduction device, the initial value of the predictive equalization coefficient pre-recorded on the optical information recording medium and the area in which the known multi-data is pre-recorded can be reproduced, and the predictive equalization coefficient It is then decided to minimize the error between the signal output of each item of the target data and the multiple data. The above and other objects, features, and advantages of the present invention will become more apparent from the following detailed description in conjunction with the accompanying drawings. [Embodiment] An embodiment of the present invention will be described below with reference to the following drawings. As described in the prior art, a conventional waveform equalization operation cannot fully eliminate waveform interference through waveform equalization. In the case where the waveform interference contains components that do not have a linear effect (as described in FIG. 18) "Regional modulation". As described with reference to Figs. 19A to 21B, the interference between codes exhibits different degrees of influence, depending on the state of the recording marks before and after the recording marks of the subject to be reproduced. To solve this problem, the detection waveform equalization operation is repeated according to the state and influence of the recording marks located before and after the main recording mark (which is a target to be reproduced). If you want to do this, you have to predict more data. However, since the reproduced signal from one of the waveform equalization before it is indistinct, an operation must be performed to increase the accuracy of data prediction. When used to increase the accuracy of data prediction, -19- (16) 1233104-a waveform equalization operation using a fixed equalization constant is performed on the reproduced signal, so that its multiple data can be accurately predicted and self-reproduced signal. In order to perform a waveform equalization operation based on the state and influence of the recording marks located before and after the main recording mark (which is a target to be reproduced), every three or more combination types must be set, Contains a record mark string before the body record mark, a body record mark, and a record mark string after the body record mark. For example, when a recording and reproduction point is located at the center of the main recording mark, the relationship between the diameter BD between the recording and reproduction point and the cell length CL is approximately BD = 2 X CL, then the recording and reproduction point is Across three record marks. Therefore, it is necessary to refer to its equalization coefficients for waveform equalization in each of the three combined patterns from the prediction data to optimize the waveform equalization conditions, and under the optimized waveform equalization conditions Perform detective waveform equalization. The above-mentioned waveform equalization conditions can be changed with the combination of the information recording device and the optical information recording medium, according to the change of the recording conditions between the optical information recording devices and the change of the recording discrimination between the recording media. Therefore, in order to effectively increase the equalization performance of the waveform according to the present invention, it is necessary to regenerate the known multiple data in advance, find the equalization condition of the predictive waveform and the equalization condition of the detection waveform, and then reproduce the result based on the result Unknown data. Fig. 1 shows a structure of a multi-data recording circuit, which is an embodiment of a multi-data recording and reproducing apparatus according to the present invention. -20- (17) 1233104 This multi-data detection circuit includes a predictive waveform equalizer 11, a predictor 1, which predicts multi-data, a detective waveform, etc. from a reproduced signal that receives the equalization of the predictive waveform. A quantizer 31 and a determiner 3 predict multiple data from a reproduced signal equalized by a detection waveform. With this circuit structure, the conditions for equalizing the waveform are obtained, and unknown data is reproduced. In this circuit structure, an initial equalization coefficient is recorded in advance on an optical information recording medium to obtain a higher efficiency of predictive waveform equalization. With this operation, it is possible to eliminate the problem of the long calculation time required due to the change of the equalization coefficient between the optical information recording media. Furthermore, in order to increase the accuracy of the prediction result, one of the waveform equalization targets is generated from a reproduced signal recorded in a multiple volume. Therefore, even if the reproduction signal fluctuates with a change such as a change in reflectance or a recording discrimination power, it is possible to predict a large amount of data having a reproduction signal fluctuation. In order to obtain the target chirp, it is effective to use the following types, taking into account the characteristics of equalization of the waveform. Figures 2 and 3 are graphs illustrating the state of convergence in the case where octal recording waveform equalization is performed. The relationship between the recording and reproduction point diameter BD and the cell length CL is roughly expressed as: BD = 2 X CL. As can be seen in Figures 2 and 3, the convergence of each item of the multiple data includes (0 0 0), (1 1 1), (2 2 2), (3 3 3), (444), (5 5 5), (666), and (777) of (\, y, 2), where X represents the multiple data before the subject record mark, y represents the multiple data about the subject record mark, and z represents the subject Record the data from the 値 -21-(18) 1233104 after the mark. Based on the above principles, the target volume can be determined simply by regenerating the types of data it records in a combination of eight types (0 0 0) to (7 7 7). In the case of octal records, the total number of combinations of (X, y, Z) is 83 = 5 1 2. Therefore, the target 値 can be obtained from a combination of only eight types, which is extremely efficient. Here, "0" represents the unit in which no record mark is recorded, and "7" represents the unit in which the most record mark is recorded. In order to effectively converge multiple data to eight targets, use the data shown in Figure 4 The waveform equalization circuit replaces the waveform equalization circuit of the conventional technique shown in FIG. The waveform equalization circuit shown in FIG. 4 has a 5-branch structure. This structure is designed to eliminate waveform interference from the two record marks before the main record mark and the two record marks after the main record mark, which has a total of four record marks. As mentioned above, when the recording and reproduction point is placed at the center of the subject record mark and spans three record marks, there is a disturbance between the subject record mark receiving codes, which are the two records from before and after the subject record mark. The impact of the mark. However, in order to increase the correction accuracy of the waveform equalizer, a 5-branch structure is more desirable, which can reduce the interference from the recording marks before and after the main recording mark. In the structure shown in Fig. 4, all the outputs of the four difference calculators 42a to 42d are ,, 0 ", when five consecutive units have the same polynomial data -22- (19) 1233104 Therefore, when the target 値When the combination of (x, y, z) is x = y = z, which is the type used to obtain the target 値, the four difference calculators 42 a to 42 d are “0”, regardless of the waveform equalization. The structure of the coefficient shown in Figure 4 acts as a waveform equalizer that converges on the target frame. At the same time, the waveform equalization circuit has four equalization coefficients, which are higher than those shown in Figure 17 The conventional circuit lacks an equalization coefficient. Therefore, this structure is more advantageous in that there are fewer calculation steps in obtaining the equalization coefficient. Next, an information record by using a multi-data detection circuit of FIG. 1 will be described. Recording and reproduction of multiple data signals, one of the reproduction and reproduction circuits. FIG. 5 is a block diagram showing the structure of an information recording and reproduction circuit using the multiple data detection circuit shown in FIG. 1. This information recording and reproduction circuit executes A data record on an optical information recording medium 50 and And includes a pickup head 51, an LD driver signal generator 52, a multi-data converter 5 3, an information data generator 5 4, a photodetector 55, an AGC controller 56, a synchronization signal unit A detection circuit 57, a sampling signal generating circuit 58, a quantitative AD converter 59, and a multi-signal memory 60. The information data generator 54 generates digital data to be recorded on the optical information recording medium 50, and more The data converter 53 converts the multiple data into octal data. If, for example, the digital data is “〇〇1 1 〇1 〇1〇”, then the data converter 53 converts the digital data Each of the three figures is converted into an octal number -23- (20) 1233104 words, and the octal data "1 52" is generated. According to the octal data, a recording pulse pattern to be recorded with the LD driver signal generator 52 is generated, and the semiconductor laser light source (LD) of the pickup head 51 is driven so that its pickup head 51 focuses the laser beam L on Rotating optical information recording medium 50. Therefore, optical recording is performed on the optical information recording medium 50. The optical information recording medium 50 may be a CD-R disc on which a pigment material such as a "write-once" material is used, or a CD-RW disc on which a rewritable phase-variable material is used. sheet. FIG. 6 shows a recording data format of the optical information recording medium 50 shown in FIG. As shown in FIG. 6, the recording data format of the optical information recording medium 50 includes unit blocks called “sectors,” such as “finding sections” and “data sections 1 to N”. Each section contains a section mark, a synchronization signal area, a single address area, and a multi-data area. The section mark indicates the beginning of the section and is a type that does not appear in the data area. (Not shown), such as "00000000777777," 〇 The synchronization signal area is formed by a repeating pattern (not shown), such as "〇07〇07〇07〇7〇7〇7〇7 ... The repeat pattern will be used as a sampling clock when multiple frames of data are quantized. The address area indicates the address of the sector, and the address is recorded as multiple frames of data. In order to reproduce a signal, laser light is focused on On the rotating optical information recording medium 50, and the reflected light is photoelectrically converted into an electric signal by the photodetector 55. The AGC controller 56 detects a segment mark from one of the reproduced signals (21) 1233104, And further detect the maximum signal of the segment mark 値 ”000 00000 "and minimum signal 値" 77777777 ". The AGC controller 56 then performs automatic gain control (AGC) to stabilize the amplitude (that is, the difference between the maximum signal 値 and the minimum signal 値). AGC is performed to correct due to The optical information recording medium 50 is caused by incorrect reflectance variation caused by the variability. The synchronization signal unit detection circuit 5 7 next detects a synchronization signal region. According to the synchronization signal, the sampling signal generation circuit 5 8 generates a Timing signal. The quantitative AD converter 59 then uses the timing signal to quantify (analog to digital conversion) the multi-signal (retaining the signal at the center of the cell with the sample hold). The quantized multi-signal is continuously recorded in the multi-signal memory 6 〇. The signal recorded in the multi-channel signal memory 60 is continuously read and input as a quantized reproduced signal in the multi-channel data detection circuit shown in Figure 1. The signal will be used for waveform equalization and multi-channel値 Decision. Next, the procedure for obtaining the predictive equalization coefficient in the multi-data detection circuit will be described. The initial step unit of FIG. 1 (ie, Spectral waveform equalizer 11 and multi-level data predictor 16) are used as one circuit structure of predictive waveform equalization. At the same time, a waveform equalizer with a circuit structure as shown in FIG. 1 is used here. The waveform equalizer shown in Figure 4 performs a calculation operation according to the following equation (7): • 25- (22) 1233104 EQ (n) = D (n) + {D⑻ 一 D (n-2)} x C0 + {D⑻- (η-l)} x C1 + {D⑻- (η + 1)} χ C2 + {D⑻- (n + 2)} xC3 .... (7) In this equation, D (n) Represents the reproduced signal of the n-th unit, EQ (n) represents the signal output 値 after the equalization operation of the predictive waveform, and C0 to C3 represent the equalization coefficients. These equalization coefficients are fixed coefficients regardless of their multiple data types. FIG. 7 is a flowchart of an operation algorithm for obtaining a predictive equalization coefficient. A known data signal has a seek section for each N consecutive data sections, as shown in the format of FIG. Find out how much data in the segment is known data. The known data is formed by data in which all combinations (5 1 2-8 = 5 04 types) are repeatedly recorded as the target 値 type. The initial equalization coefficient is pre-recorded in an area separated from the data recording area, and information will be obtained by reproducing the separated area. However, if the waveform equalization operation is not optimized, accurate multi-data cannot be detected. Therefore, it is more desirable to record its initial equalization coefficient as binary information so that its initialization coefficient can be accurately obtained. The determination of the predictive equalization coefficient is repeated until the error with respect to the target signal becomes equal to a predetermined value (T0) or less. To calculate the error, it is necessary to obtain information about the distribution of the known data. Information about the configuration of the known data is recorded in a memory (not shown) in the information recording and reproduction device shown in FIG. 5 and is used to record and reproduce the requested information. -26- (23) 1233104 material. The multi-signal memory 60 shown in FIG. 5 needs a data capacity equal to or larger than its size corresponding to the time period required to determine the equalization coefficient. With a sufficiently large data capacity, the unknown multiple data that will be reproduced during the predictive equalization coefficient determining operation and subsequent detection equalization coefficient determining operations can be temporarily stored, and data processing can be continuously performed . After the predictive equalization coefficient is determined, the detection equalization coefficient is started to be obtained upon receiving the notification that the predictive equalization coefficient decision has been completed. At the time of reproduction of the known data, the obtained results (ie, the determined predictive equalization coefficients CO, Cl, C2, and C3) are set to the predictive equalization coefficient calculator 12 shown in FIG. 1, And then accept the waveform equalization. The predictive data used in the equalization of the detection waveform is the multi-data obtained by the multi-data predictor 16. This multi-data predictor i 6 transforms the signal that it accepts the prediction waveform equalization. At this point, the threshold is determined, and the target data of the self-test are determined. The target frame is recorded in a predictive threshold frame generator 17 and used as a threshold frame information by the multiple frame data predictor 16. The results (ie, the predictive equalization coefficients C 0, C 1, C 2, and C 3, and the threshold based on the number of targets (data)) are updated every time the calculation shown in Figure 6 When one of the fetched segments is reproduced, it is used to reproduce unknown multi-data. Next, a procedure for obtaining the detection equalization coefficient in the multi-data detection circuit will be described. The latter half of FIG. 1 (i.e., the detection waveform equalizer 31 and the multi-data detector 34) are used as a circuit structure for detection waveform equalization. At the same time, a waveform -27- (24) 1233104 equalizer with a circuit structure as shown in Fig. 4 is used here. The waveform equalizer shown in FIG. 4 performs a calculation operation according to the following equation (8): EQ, (n) = D (n) + {D (n)-D (n-2)} x CO (I , J, K) + {D (n)-(η-l)} x Cl (I small K) + {D (n)-(η + l)} x C2 (I, J, K) + {D⑻ -(η + 2)} x C3 (I, J, K) ..... (8) In this equation, D (n) represents the regeneration signal of the nth unit, and EQ '(η) represents the detection waveform. The signal output after the equalization operation is 値. Here, CO (I'J, K) to C3 (I, J, K) represent equalization coefficients corresponding to each S!] Type (I, J, K). In addition, I represents the item (n-i) of the multi-data, f is the n item of the multi-data, and K represents the (η + 1) item of the multi-data. FIG. 8 is a flowchart of an operation algorithm for obtaining a detection equalization coefficient. Each known data signal has a format as shown in Figure 6, which is the same as a predictive waveform equalization. The known data is formed by data in which all combinations (5 1 2-8 = 5 04) are repeatedly recorded as the target 値 pattern. In order to reproduce the known data and calculate the waveform equalization coefficients so that 値 of EQ ′ (η) converges to the target 値, then the result of at least four consecutive units (for example, EQ (n-2), EQ (nl), EQ ( n), and the result of EQ (n + l)) need to be calculated according to equation (7). Therefore, it is necessary to find the synchronization equation and obtain the optimal waveform equalization coefficient for each 5 1 2 type (xyz). This involves a lot of calculations and requires extremely long seek times. -28- (25) 1233104 However, the error of the target 値 generated during the equalization of the predictive waveform is directly proportional to the equalization coefficients CO, Cl, C2 determined by the equalization of the predictive waveform of equation (7). , And C3. Therefore, when the correlation error of the corresponding target 时 when the η term of the multi- 値 data is reproduced is 5 (η), the following equations (9) to (12) can be established. The error from {D (n) -D (n-2)} is 0c C0 / {| C0 | + | Cl | + | C2 | + | C3 |} xa (n)… (9) from {0 ( 11) -0 (11-1)} error 〇〇 (: 1 / {| C0 | + | Cl | + | C2 | + | C3 |} xa (n). &Quot; (10) From ⑺㈠ 丨 ^ ⑶ +: ^ 丨 The resulting error ^ 匸 二 / {| C0 | + | Cl | + | C2 | + | C3 |} xa (n)… (1 1) from {D (n) -D (n + 2 )} The resulting error is 0c C3 / {| C0 | + | Cl | + | C2 | + | C3 |} xa (n)-(12) From the above equations (9) to (l2), it is relevant to when The error 3 (η) of the corresponding target 时 when the η term of the multi-data is reproduced. In addition, the optimal equalization coefficients C0 (I, J, κ), C1 (I, J, K), C2 (I, J, K), and C3 (I, J, K) can be calculated by the following equations (1 3) to (16) using the errors 5 (η) and Predictive equalization coefficients CO, Cl, C2, and C3. It should be noted that I represents the (η-1) term of the multivariate data, J represents the η term of the multivariate data, and K represents the (η) of the multivariate data. + 1) term. -29-(26) 1233104 C0 (I, J, K) = C0-a (n) x SOx {D (n) -D (n-2)} x G… (13) C1 ( I, J, K) = Cl-a (n) x Six {D (n) -D (nl)} x G… (14) C2 (I, J, K) = C2-a (n) x S2x {D (n) -D (n + l)} x G ··· (15) C3 (I, J, K) = C3-a (n) x S3x {D (n) -D (n + 2)} x G ··· (16) In the above equation, G represents the convergence gain. Meanwhile, SO, SI, S2, and S3 are represented by the following equations (17) to 20), where IPI represents the absolute 値 of P. so = co / {l CO |-H Cl 1 + | C2 1 + 1 C3 1}. .. (17) S 1 = Cl / { l CO |-M ci | + | C2 1 + 1 C3 1 • (18) S2 = C2 / {1 CO 1 1 M ci | + | C2 1 + 1 C3 1 • (19) S3 = C3 / {| CO 1 HH ci | + | C2 1 + 1 C3 1} .. (20) 値 C0 (I, J, K), C1 (I, J, K), C2 (I, J, K), and C3 ( I, J, K) can be determined from 6 (η) through operations according to equations (13) to (16). Therefore, no synchronization equation is required. In this way, the optimal equalization coefficients can be automatically determined from the error 5 (η) of the η term of the multi-data and the predictive equalization coefficients CO, Cl, C2, and C3. Therefore, the calculation time of the detection equalization coefficient can be significantly reduced. The determination of the detection equalization coefficient is repeated until the error with respect to the corresponding target becomes equal to a predetermined value (T0 ') or less. During the reproduction of unknown data, determine (ie, the determined detectability -30- (27) 1233104 equalization coefficients C0 (I, J, K), Cl (I, J, K), C2 (I, J, κ), and C3 (l, j, K), and convergence gain G) are set to the detectable equalization coefficient and convergence gain calculator 32 in FIG. 1, and then waveform equalization is performed. FIG. 9 shows a table of equalization coefficients used to detect waveform equalization. The detection equalization coefficient is set and stored in the detection equalization coefficient and reception gain g calculator 32, based on the reproduced known data. The setting of the 'detective equalization coefficient' as shown in Fig. 9 is performed for each unit of three consecutive patterns. Based on the multi-unit data obtained by the multi-unit data predictor 16, the optimal equalization coefficient and convergence gain are read from the detectable equalization coefficient and convergence gain calculator 32. Thereafter, the detection waveform equalization is performed, and then the multi-data detector 34 performs multi-data determination. In order to determine multiple data with high accuracy, it is more desirable to use the result of the multiple signal calculated when the detection waveform is equalized, which is better than using the target data obtained from the predictive threshold. Therefore, when the known data is reproduced and obtained, the signal distribution at each multi-level receives its statistical processing based on the results of the multi-level data and the reproduced signal, and the obtained signal is used as the next multi-level decision. Threshold 値. Therefore, the above-mentioned threshold calculation operation is performed using the multiple threshold signal memories 36 and the statistics processor 37 shown in FIG. 1, and the obtained threshold threshold is set to the detection threshold threshold generator 3 5. Using these threshold chirps, the multiple chirp signals after the detection waveform is equalized can be accurately determined. Find the result (that is, the determined detectable equalization coefficients C0 (I, J, K), C 1 -31-(28) 1233104
(I，J，K)、C2 (I ’ J，K)、及 C3 (I，J，κ)、收斂增益 G 、以及多値資料決定臨限値）被更新於每次圖6中所示的 求取區段之一被再生時，而更新的求取結果被使用以再生 未知的多値資料。 設定多値資料決定臨限値之技術係由預測性波形等化 器1 1所利用，以致其可進一步增加決定預測性資料之準 確性。 雖然未顯示於圖形中，一統計操作及一多値記憶體被 加至圖1中所示之預測性臨限値產生器1 7，以致其臨限 値可被求取以預測性波形等化器1 1。 於未知資料之多値決定時，檢測性波形等化被執行於 透過預測性波形等化之多値資料的預測以後，而因此，需 要一時間延遲。 因此，在暫時儲存未知的多値資料於記憶體緩衝器之 後，預測性波形等化及預測性多値決定被完成，且未知資 料被再生自記憶體緩衝器。檢測性波形等化及多値決定被 接著執行。 接下來，將描述多値資料檢測電路中再生未知資料之 一範例操作。 圖1 0係多値資料檢測電路中再生未知資料之操作演 算法的流程圖。 根據其透過求取預測性等化係數及檢測性等化係數之 程序所獲得的結果，多値決定被執行於未知的資料串。如 此一來，圖6中所示之格式中的求取區段中之多値資料（ -32- (29) 1233104 其具有一求取區段於每一 N連續未知資料區段）被使用 爲已知資料。因此，根據求取區段之再生的結果，則多値 決定被執行於未知的資料區段。求取區段之***頻率（一 求取區段被***每一 N未知的資料區段）被決定自光學 資訊記錄媒體5 0之信號位準波動的頻率特性。 假如光學資訊記錄媒體50之每一周造成一信號位準 波動，舉例而言，則至少四個求取區段需被***於一周， 以追隨波動頻率。假如每一徑向區域中造成一信號位準波 動（亦即，於一周中信號位準波動兩次），則一求取區段 應被***於各徑向區域中之每數毫米中，以追隨波動。 假如信號位準波動於光學資訊記錄媒體5 0之間，則 求取區段應僅被再生自最內周圍區域及最外周圍區域，其 係分離自資料記錄區域。此一格式係不同於圖6中所示之 格式。於此情況下，求取區段未被***資料區域中，且記 錄及再生係根據求取結果而被執行。 信號位準波動可隨光學資訊記錄媒體及資訊記錄與再 生裝置之組合而改變。因此，每次一光學資訊記錄媒體被 設定至資訊記錄及再生裝置時（亦即，每次新的光學資訊 被記錄時），則求取區段僅被記錄及再生於最內周圍區域 及最外周圍區域，而求取係透過這些求取區段而被執行。 於此情況下，求取區段未被***資料區域中。因此，資料 區域中之記錄及再生係根據求取結果而被執行，以致其信 號位準波動可被恆定地追隨。 接下來，將描述圖6中所示之格式下執行以N (二 -33- (30) 1233104 1 2 8 )區段中一次之求取頻率的求取程序之結果。亦將描 述透過此一求取程序之未知記錄資料的評估結果。每一區 段中之多値資料包含2048個單元（2048x3位元；八進記 錄）。 圖1 1顯示在波形等化之前於個別多値位準上的再生 信號之分佈。圖1 2顯示在預測性波形等化之後於個別多 値位準上的再生信號之分佈。圖1 3顯示在檢測性波形等 化之後於個別多値位準上的再生信號之分佈。 透過此實施例之波形等化，則個別多値位準上之再生 信號的分佈可被顯著地增進，且再生信號有效地收斂。於 此，預測性等化係數CO，Cl，C2,及C3個別爲0.00， 0.15，0.1 8,及- 0.01。同時，檢測性波形等化中之收斂增 益G爲3 8。 此處所使用之拾取頭的光學系統具有650 nm之記錄 及再生波長λ、0.65 NA之物鏡、及約0.8微米之光束點 直徑BD。單元長度係0.46微米，而記錄密度爲6.52位 元/微米。相較於其可被記錄及再生以相同拾取頭之可重 寫光學資訊記錄媒體（以二元記錄之3.75位元/微米的 DVD )，此實施例中之記錄密度約更高1 . 7倍。以多値信 號之最大値及最小値爲” 1 ”，則每一多値位準上之分佈的 評估係由標準偏差之平均値σ avg所界定。此實施例之波 形等化的效果被相互比較，使用多値決定中之單元單位的 平均値σ avg及誤差率Err。其比較結果如下： (31) 1233104 在波形等化之前： σ avg = 6.5%，Err = 20% 至 40% 在預測性波形等化之後： σ avg = 2.6%，Err = 2% 至 3% 在檢測性波形等化之後： σ avg = 1.6%，Err = 0%至 0.05% 上述結果證實其平均値σ avg減少至1/4而誤差率 Err減少至1 /400或更低，其係由於此實施例之效果。 以此方式，此實施例之多値資料檢測電路預測其位於 待被再生主體記錄標記之前及之後的記錄標記之狀態及影 響程度，當從再生信號再生資訊時，該等再生信號係透過 記錄標記之區域的調變而使其位準被多値化。根據所預測 的影響程度，最佳檢測性波形等化被最終地執行於每一型 態。因此，碼之間的干擾可透過波形等化而被有效地消除 〇 同時，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，等化係數條件被 決定於其包含主體記錄標記前之記錄標記串的已知資料値 、主體記錄標記之已知資料値、及主體記錄標記後之記錄 標記串的已知資料値的三個或更多連續記錄標記之各組合 型態。因此，能夠以高精確度求取其藉由一波形等化操作 以消除碼之間干擾的條件。 再者，當欲從再生信號（其係透過記錄標記之區域的 -35- (32) 1233104 調變而使其位準被多値化）再生資訊時，決定臨限値係從 預測性波形等化及檢測性波形等化後之已知資料的多値輸 出被決定。因此，能夠以高精確度設定多値決定中之臨限 値。 同時’當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，則利用上述等化 係數求取技術及臨限値求取技術。因此，能夠以高精確度 透過一波形等化操作而消除碼之間的干擾，並執行精確的 多値決定於未知資料上。 再者，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，記錄資料之各項 的信號輸出之目標値係在藉由再生其含有相同多値資料之 三個或更多記錄標記串而獲得的波形等化之前的信號輸出 値。因此，可於波形方程式中獲得有效的收斂。 同時，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，上述目標値被設 定爲收斂目標。因此，可於波形方程式中獲得有效的收斂 ，而等化係數之數目被減一，相較於習知的電路結構（參 見圖1 7 )。因此，求取等化係數之程序中的計算量可被 有利地減少。 再者，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，則包含多値資料 之三個或更多連續項的所有組合之資料串被重複地記錄以 形成多値資料串，且記錄區域被週期性地形成於一分離自 -36- (33) 1233104 其中記錄有未知多値資料之資料區域的區域中，以致其光 學資訊記錄媒體之信號位準波動可被恆定地追隨。因此， 碼之間的干擾可依據信號位準波動而被消除，且精確的多 値決定可被執行。 同時，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，其指示波形等化 之收斂目標的型態串被記錄爲求取資訊，而求取區域被週 期性地形成於一分離自其中記錄有未知多値資料之資料區 域的區域中，以致其光學資訊記錄媒體之信號位準波動可 被恆定地追隨。因此，碼之間的干擾可被消除，且準確的 多値決定可被執行。 再者，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，相關於其產生於 預測性波形等化中之目標値的誤差被假設爲直接正比於其 透過根據方程式（7)之預測性波形等化而決定的等化係數 CO，Cl，C2,及C3，且被幾乎均勻地分佈。檢測性波形等 化係數係依此假設而被計算。因此，最佳等化係數可被自 動地決定自第i項多値資料之誤差5 (η)及預測性等化係數 CO，Cl，C2,及C3。因此，用以計算檢測性等化係數所需 的時間可被顯著地縮短。 同時，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，相關於其產生於 預測性波形等化中之目標値的誤差被假設爲直接正比於其 透過根據方程式（7)之預測性波形等化而決定的等化係數 (34) 1233104 CO ’ Cl ’ C2,及C3，且被幾乎均勻地分佈。檢測性波形等 化係數係依此假設而被計算，且收斂增益G被決定以致 其相關於目標値之誤差可被減至最小。因此，最佳收斂增 益G可被自動地決定自第丨項多値資料之誤差5 (n)及預 測性等化係數CO，Cl，C2,及C3。因此，檢測性等化係 數所需的計算時間可被顯著地縮短。 再1者’當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，初始預測性等化 係數被記錄於一分離自資料記錄區域之區域中，且資訊係 藉由再生分離的區域而被獲得。因此，預測性等化求取所 需的計算時間可被顯著地縮短。 同時’當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，初始預測性等化 係數被記錄於一分離自資料記錄區域之區域中。初始預測 性等化係數被記錄爲二元資訊，以致其初始係數可被準確 地獲得。因此，預測性等化求取所需的計算時間可被顯著 地縮短。 接下來，將描述本發明之另一實施例。 上述多値資料檢測電路有效地消除碼之間的非線性干 擾成分。以此多値資料檢測電路，則等化係數之計算可被 大大地簡化。 然而，假如預測性決定結果中的誤差率爲高時，則由 於波形等化所致之錯誤成分可能增加且碼之間的千擾成分 可能無法被有效地消除。 -38- (35) 1233104 以圖1中所示之多値資料檢測電路，碼之間的干擾成 分會增加，因爲單元長度縮短且記錄密度增加。因此，預 測性決定値之誤差率以及“固定等化器及臨限値確認，，之組 合變得較高。 於此情況下，檢測性波形等化係根據具有高誤差率之 預測性決定結果而被執行。結果，波形等化被執行以不正 確的等化係數，且檢測性波形等化之後的決定結果無法被 充分地減小。 如此一情況之範例，記錄及再生之結果將被顯示於下 。於此記錄及再生操作中，拾取頭之光學系統具有650 nm之記錄及再生波長λ、0.65 NA之物鏡、及約0.8微米 之光束點直徑。單元長度係0.40微米，而記錄密度爲7.0 位元/微米。 預測性決定結果： 誤差率1 0 . 1 % 圖1中所示之電路的檢測性決定結果： 誤差率7.6 % 從以上結果可知，先前實施例具有增進約3 0 %誤差 率之效果。 爲了進一步增加其增進檢測性能之效果’則於此實施 例中利用一種具有如圖1 4中所示之電路結構的多値資料 檢測電路。於此電路結構中，多値決定演算法被增進如下 〇 於此多値決定演算法中’先前實施例之檢測性決定結 -39- (36) 1233104 果被使用爲預測結果，且根據預測結果之先前實施例的檢 測性波形等化被重複直到多値資料之決定結果充分地收斂 〇 於其利用5分支結構之波形等化器的先前實施例中， 最佳波形等化係數被決定自待被再生之主體記錄標記的已 知資料値、主體記錄標記前之記錄標記串的已知資料値、 及主體記錄標記後之記錄標記串的已知資料値之組合（參 見圖9 )。另一方面，於此實施例中，最佳波形等化係數 被決定自主體記錄標記前之記錄標記串的已知資料値及主 體記錄標記後之記錄標記串的已知資料値之組合，其係使 用圖1 5中所示之一等化係數表列。 此實施例之波形等化操作係更有別於先前實施例之波 形等化操作。然而，於此實施例中，波形等化係數被重複 直到多値資料之決定結果充分地收斂。以此方式，即使藉 由圖1之“固定等化器及臨限値確認”所獲得的預測性決定 値之誤差率爲高，其檢測之決定結果的誤差率仍可被降低 〇 藉由一種利用圖1中所示之多値資料檢測電路的資訊 記錄再生電路而透過多値資料信號之記錄及再生所再生的 再生信號被輸入圖1 4中所示之檢測電路。接著利用此多 値資料檢測電路以執行如先前實施例中之相同操作來求取 預測性等化係數。 接下來，將描述利用此多値資料檢測電路以求取檢測 性等化係數之程序。下列描述將省略先前實施例中對相同 -40- (37) 1233104 程序之解釋。 圖1 4中所示透過初始檢測性波形等化而計算之多値 信號的求取結果係相同於串聯配置之後者多値決定電路中 所使用的求取結果。 於未知資料之多値決定中，檢測性波形等化被執行於 多値資料透過預測性波形等化而被預測之後，而因此，需 要一時間延遲。 有鑑於此，在未知多値資料被暫時地儲存於記憶體緩 衝器中之後，預測性波形等化及預測性多値決定便完成。 未知資料被接著再生自記憶體緩衝器，且檢測性波形等化 及多値決定被依序地執行。 以上述之相同方式，個別的決定結果被接收，且多値 決定被執行於其使用相同等化係數以執行波形等化操作之 檢測器 102、104、及106中。因此，從未知多値資料被 暫時儲存於記憶體緩衝器中起已經過了預測性多値決定所 需的預定時間週期以後，預測性波形等化及預測性多値決 定便完成。未知資料被接著再生自記憶體緩衝器，且多値 決定被重複地執行。(I, J, K), C2 (I'J, K), and C3 (I, J, κ), convergence gain G, and multi-data data threshold (决定) are updated each time as shown in FIG. 6 When one of the evaluation sections of is reproduced, the updated evaluation results are used to reproduce unknown multi-data. The technique of setting multiple data to determine the threshold is used by the predictive waveform equalizer 11 so that it can further increase the accuracy of determining predictive data. Although not shown in the figure, a statistical operation and a multi-frame memory are added to the predictive threshold generator 17 shown in FIG. 1 so that its threshold can be calculated to equalize the predictive waveform.器 1 1. When the amount of unknown data is determined, the detection waveform equalization is performed after the prediction of the amount of data through the predictive waveform equalization. Therefore, a time delay is required. Therefore, after temporarily storing the unknown multiple data in the memory buffer, the equalization of the predictive waveform and the predictive multiple decision are completed, and the unknown data is regenerated from the memory buffer. Detect waveform equalization and multiple decision are then performed. Next, an example operation of reproducing unknown data in the multi-data detection circuit will be described. Fig. 10 is a flowchart of an operation algorithm for reproducing unknown data in a multi-data detection circuit. Based on the results obtained through its procedure of obtaining predictive equalization coefficients and detective equalization coefficients, it was decided to be executed on unknown data strings. As a result, the multiple pieces of data (-32- (29) 1233104 in the seeking section in the format shown in FIG. 6 having a seeking section in every N consecutive unknown data sections) are used as Known information. Therefore, based on the results of the regeneration of the obtained segments, it is often decided to be executed on unknown data segments. The insertion frequency of the determination section (one determination section is inserted into each N unknown data section) is determined from the frequency characteristics of the signal level fluctuation of the optical information recording medium 50. If the optical information recording medium 50 causes a signal level fluctuation every week, for example, at least four seeking sections need to be inserted in a week to follow the fluctuation frequency. If a signal level fluctuation is caused in each radial area (that is, the signal level fluctuates twice in a week), a finding section should be inserted in every few millimeters in each radial area to Follow the volatility. If the signal level fluctuates between 50 in the optical information recording medium, the seeking section should be reproduced only from the innermost peripheral area and the outermost peripheral area, which is separated from the data recording area. This format is different from the format shown in FIG. In this case, the seeking section is not inserted into the data area, and recording and reproduction are performed based on the finding result. The signal level fluctuation can be changed by the optical information recording medium and the combination of the information recording and reproduction device. Therefore, each time an optical information recording medium is set to the information recording and reproducing device (that is, each time new optical information is recorded), the seeking section is recorded and reproduced only in the innermost surrounding area and the most In the outer surrounding area, the evaluation is performed through these evaluation sections. In this case, the seeking section is not inserted into the data area. Therefore, the recording and reproduction in the data area are performed based on the result of the determination, so that fluctuations in its signal level can be constantly followed. Next, a description will be given of a result of performing an obtaining procedure for obtaining a frequency in the N (two -33- (30) 1233104 1 2 8) section in the format shown in FIG. 6. The results of the evaluation of unknown recorded data through this process will also be described. The data in each section contains 2048 units (2048x3 bits; Octal records). Figure 11 shows the distribution of reproduced signals at individual multi-levels before the waveform is equalized. Figure 12 shows the distribution of the reproduced signals at the individual multi-levels after the prediction waveform is equalized. Figure 13 shows the distribution of the reproduced signal at the individual multi-levels after the detection waveform is equalized. By equalizing the waveforms of this embodiment, the distribution of the reproduced signals at individual multi-levels can be significantly improved, and the reproduced signals effectively converge. Here, the predictive equalization coefficients CO, Cl, C2, and C3 are individually 0.00, 0.15, 0.1 8, and -0.01. At the same time, the convergence gain G in the detection waveform equalization is 38. The optical system of the pickup head used here has a recording and reproduction wavelength λ of 650 nm, an objective lens of 0.65 NA, and a beam spot diameter BD of about 0.8 m. The cell length is 0.46 m, and the recording density is 6.52 bits / m. Compared with a rewritable optical information recording medium (a DVD recorded at 3.75 bits / micron) which can be recorded and reproduced with the same pickup, the recording density in this embodiment is about 1.7 times higher. . Taking the maximum and minimum values of the multi-signal as "1", the evaluation of the distribution at each multi-level is defined by the mean 値 σ avg of the standard deviation. The effects of the waveform equalization in this embodiment are compared with each other, using the average 値 σ avg and the error rate Err of the unit units in the multiple determination. The comparison results are as follows: (31) 1233104 Before waveform equalization: σ avg = 6.5%, Err = 20% to 40% After predictive waveform equalization: σ avg = 2.6%, Err = 2% to 3% in After the detection waveform is equalized: σ avg = 1.6%, Err = 0% to 0.05%. The above results confirm that the average 値 σ avg is reduced to 1/4 and the error rate Err is reduced to 1/400 or less. Effect of the embodiment. In this way, the multiple data detection circuit of this embodiment predicts the state and influence of the recording marks before and after the recording marks of the subject to be reproduced. When information is reproduced from the reproduced signals, the reproduced signals are transmitted through the recorded marks. The adjustment of the area makes its level multiplied. Based on the predicted degree of influence, the best detectable waveform equalization is ultimately performed for each type. Therefore, the interference between codes can be effectively eliminated by equalizing the waveform. At the same time, when information is to be reproduced from a reproduced signal (which is multiplied by the modulation of the area of the recorded mark), The equalization factor condition is determined by three or more of the known data of the record mark string before the main record mark, the known data of the main record mark, and the known data of the record mark string after the main record mark. More continuous recording marks of each combination type. Therefore, it is possible to obtain a condition with high accuracy to eliminate interference between codes by a waveform equalization operation. Furthermore, when the information is to be reproduced from a reproduction signal (which is -35- (32) 1233104 modulated through the area where the mark is recorded), the threshold is determined to be from a predictive waveform, etc. The multiple outputs of known data after equalization and detection waveform equalization are determined. Therefore, it is possible to set the threshold in the multi-decision decision with high accuracy. At the same time, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recording mark), the above-mentioned equalization coefficient determination technique and threshold detection technique are used. Therefore, it is possible to eliminate interference between codes by performing a waveform equalization operation with high accuracy, and to perform accurate multiplication depends on unknown data. Furthermore, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recording mark), the goal of the signal output of each item of recorded data is to reproduce it The signal output before equalizing the waveform obtained from three or more record mark strings containing the same amount of data is output. Therefore, effective convergence can be obtained in the waveform equation. At the same time, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recording mark), the above target 値 is set as a convergence target. Therefore, effective convergence can be obtained in the waveform equation, and the number of equalization coefficients is reduced by one, compared with the conventional circuit structure (see Fig. 17). Therefore, the calculation amount in the procedure for obtaining the equalization coefficient can be advantageously reduced. Furthermore, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recorded mark), it contains all combinations of three or more consecutive items of multi-data The data string is repeatedly recorded to form a multi-data string, and the recording area is periodically formed in an area separated from the data area in which unknown multi-data is recorded, so that its optical information Signal level fluctuations of the recording medium can be constantly followed. As a result, interference between codes can be eliminated based on signal level fluctuations, and accurate multiple decisions can be performed. At the same time, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recording mark), the pattern string indicating the convergence target of the waveform equalization is recorded as the information seeking The seeking area is periodically formed in an area separated from the data area in which unknown data is recorded, so that the signal level fluctuations of its optical information recording medium can be constantly followed. Therefore, interference between codes can be eliminated, and accurate multiple decisions can be performed. Furthermore, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recorded mark), errors related to the target 値 generated in the equalization of the predictive waveform are detected. It is assumed that it is directly proportional to the equalization coefficients CO, Cl, C2, and C3 determined by equalization of the predictive waveform according to equation (7), and is distributed almost uniformly. The detection waveform equalization coefficient is calculated based on this assumption. Therefore, the optimal equalization coefficient can be automatically determined from the error 5 (η) of the i-th multivariate data and the predictive equalization coefficients CO, Cl, C2, and C3. Therefore, the time required to calculate the detectable equalization coefficient can be significantly reduced. At the same time, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recorded mark), an error related to the target 产生 generated in the equalization of the predictive waveform is assumed. It is directly proportional to its equalization coefficient (34) 1233104 CO 'Cl' C2, and C3, which is determined by equalizing the predictive waveform according to equation (7), and is distributed almost uniformly. The detection waveform equalization coefficient is calculated based on this assumption, and the convergence gain G is determined so that the error related to the target chirp can be minimized. Therefore, the optimal convergence gain G can be automatically determined from the error 5 (n) of the multivariate data and the predictive equalization coefficients CO, Cl, C2, and C3. Therefore, the calculation time required for the detection equalization coefficient can be shortened significantly. Further, when the information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recording mark), the initial predictive equalization coefficient is recorded in a separate area from the data recording area. Area, and information is obtained by regenerating separate areas. As a result, the computational time required for predictive equalization can be significantly reduced. At the same time, when information is to be reproduced from a reproduction signal (which is multiplied by adjusting the level of the recorded mark area), the initial predictive equalization coefficient is recorded in an area separated from the data recording area . The initial predictive equalization coefficient is recorded as binary information, so that its initial coefficient can be accurately obtained. As a result, the computational time required for predictive equalization can be significantly reduced. Next, another embodiment of the present invention will be described. The above-mentioned multiple data detection circuit effectively eliminates non-linear interference components between codes. With this multiple data detection circuit, the calculation of the equalization coefficient can be greatly simplified. However, if the error rate in the predictive decision result is high, the error component due to waveform equalization may increase and the perturbation component between codes may not be effectively eliminated. -38- (35) 1233104 With the multiple data detection circuit shown in Figure 1, the interference component between codes increases because the unit length is shortened and the recording density is increased. Therefore, the combination of the error rate of the predictive decision frame and the "fixed equalizer and threshold confirmation" becomes higher. In this case, the detection waveform equalization is based on the predictive determination result with a high error rate. As a result, the waveform equalization is performed with incorrect equalization coefficients, and the decision result after the detection of the waveform equalization cannot be sufficiently reduced. As an example of this situation, the recording and reproduction results will be displayed Below, in this recording and reproducing operation, the optical system of the pickup head has a recording and reproducing wavelength λ of 650 nm, an objective lens of 0.65 NA, and a beam spot diameter of about 0.8 microns. The unit length is 0.40 microns, and the recording density is 7.0 bits / micron. Predictive decision result: Error rate 10. 1% Detectable decision result of the circuit shown in Figure 1: Error rate 7.6% From the above results, it can be seen that the previous embodiment has an error of about 30%. In order to further increase the effect of improving the detection performance ', in this embodiment, a multiple data detection circuit having a circuit structure as shown in FIG. 14 is used. In this circuit structure, the multi-decision algorithm is enhanced as follows. In this multi-decision algorithm, the detection result of the previous embodiment -39- (36) 1233104 is used as the prediction result, and according to the prediction result, The detection waveform equalization of the previous embodiment is repeated until the decision result of the multiple data is sufficiently converged. In the previous embodiment of the waveform equalizer using the 5-branch structure, the optimal waveform equalization coefficient is determined since it is to be determined. The combination of the known data of the main body recording mark 値, the known data of the recording mark string before the main body recording mark, and the known data of the recording mark string after the main body recording mark (see Figure 9). In this embodiment, the optimal waveform equalization coefficient is determined from the combination of the known data of the recording mark string before the main recording mark and the known data of the recording mark string after the main recording mark. One of the equalization coefficients shown in Table 15. The waveform equalization operation of this embodiment is more different from the waveform equalization operation of the previous embodiment. However, in this embodiment, the waveform The coefficients are repeated until the decision result of the multiple data is sufficiently converged. In this way, even if the error rate of the predictive decision 値 obtained by the “fixed equalizer and threshold 値 confirmation” in FIG. 1 is high, its The error rate of the determination result of the detection can still be reduced. The reproduction signal reproduced by recording and reproducing the multiple data signals is inputted by an information recording and reproduction circuit using the multiple data detection circuit shown in FIG. 1. The detection circuit shown in FIG. 4 is then used to obtain the predictive equalization coefficient by performing the same operation as in the previous embodiment using this multi-data detection circuit. Next, the use of this multi-data detection circuit to obtain The procedure of taking the detection equalization coefficient. The following description will omit the explanation of the same -40- (37) 1233104 procedure in the previous embodiment. The results of the multi-signal calculations shown by the equalization of the initial detection waveform shown in Figure 14 are the same as the multi-signal determination results used in the series configuration. In the determination of the amount of unknown data, the detection of waveform equalization is performed after the data is predicted through the prediction of waveform equalization. Therefore, a time delay is required. In view of this, after the unknown multiple data is temporarily stored in the memory buffer, the equalization of the predictive waveform and the decision of the predictive multiple are completed. The unknown data is then reproduced from the memory buffer, and the detection waveform is equalized and multiple decisions are executed sequentially. In the same manner as described above, individual decision results are received, and multiple decisions are performed in the detectors 102, 104, and 106 which use the same equalization coefficient to perform the waveform equalization operation. Therefore, after the predetermined time period required for the predictive multi-decision determination has been temporarily stored in the memory buffer, the prediction waveform equalization and the predictive multi-decision determination are completed. Unknown data is then regenerated from the memory buffer, and multiple decisions are repeatedly performed.
於5分支之波形等化操作中，所求取的檢測性等化係 數 CO (I，J，K)、Cl (I，J，K)、C2 (I，J，K)、及 C3 (I ，J，K)形成一含有512 x 4 = 2048個係數之表列於待被 再生之主體記錄標記的已知資料値、主體記錄標記前之記 錄標記串的已知資料値、及主體記錄標記後之記錄標記串 的已知資料値之每一組合。 -41 - (38) 1233104 爲了將此表列轉變爲此實施例中所使用之等化係數表 % &下計算操作需被執行於主體記錄標記前之記錄標記 串的S知資料値及主體記錄標記後之記錄標記串的已知資 料値之每一組合。 其中X代表主體記錄標記前之記錄標記串的一已知資 # ® ’ y = o係主體記錄標記後之記錄標記串的一已知資料 値’ ffi! a 1，代表等化係數，等化係數被表示爲··Detected equalization coefficients CO (I, J, K), Cl (I, J, K), C2 (I, J, K), and C3 (I , J, K) forms a table containing 512 x 4 = 2048 coefficients known information of the subject record mark to be reproduced, known data of the record tag string before the subject record mark, and the subject record mark Each combination of known data of the subsequent record mark string. -41-(38) 1233104 In order to convert this table into the equalization coefficient table% & used in this embodiment, the calculation operation needs to be performed before the main record mark. Each combination of the known data of the record mark string after the record mark. Where X represents a known data of the record mark string before the main record mark # ® 'y = o is a known data of the record mark string after the main record mark 値' ffi! A 1, which represents the equalization coefficient, equalization The coefficient is expressed as ...
(al + a2 + a3+a4 + a5 + a6 + a7 + a8)/8 其將各値平均而不管待被再生之主體記錄標記的已知 値。 結果，此實施例中所使用之等化係數表列變爲一含有 6 4x4 = 2 56個係數之表列。求取的等化係數於串聯配置之 所有多値決定電路中均爲相同的。 接下來，將描述以此多値資料檢測電路所執行之未知 資料的再生操作之一範例。 於此實施例之多値資料檢測電路中，檢測性波形等化 及決定需在多値資料之決定結果收斂以前被重複三次。 圖1 6係以此多値資料檢測電路來再生未知資料之一 操作演算法的流程圖。 根據預測性等化係數之求取結果以及檢測性等化係數 之求取結果，多値決定被執行於未知的資料串。於此’以 圖6所示之格式的求取區段中之多値資料被使用爲已知資 -42- (39) 1233104 料。於圖6所示之格式中，一求取區段被***每一 N個 連續的未知資料區段。因此，根據求取區段之再生的結果 ，未知資料區段之多値決定被執行。求取區段之***頻率 (一求取區段被***每一 N個未知資料區段中）係由光 學資訊記錄媒體之信號位準波動的頻率特性所決定。 假如信號位準波動被產生於光學資訊記錄媒體5 0之 每一周，舉例而言，則至少四個求取區段需被***於一周 內，以追隨波動頻率。假如一信號位準波動被產生於每一 徑向區域，則一求取區段應被***於每一徑向區域之每一 數微米內，以追隨波動。 當欲更新求取結果時，最好是加入先前的求取結果以 及新近再生的求取資訊至求取結果，以致其求取結果不會 由於缺陷等之突然改變而被不利地影響。 於其中求取結果係透過對1 6個求取區段之統計操作 而被計算的情況下，舉例而言，則” 1 6個求取區段之第 一求取區段的資訊”被取代以”新近再生的求取區段資訊，， ’於統計操作中。以此方式，最近的求取結果可被反應於 更新的求取結果中，而來自”由於缺陷之突然改變，，的不 利影響可被避免於求取程序中。 假如信號位準波動被產生於每次設定一新的光學資訊 記錄媒體5 0時，則求取區段應被記錄入且僅被再生自最 內周圍區域及最外周圍區域，其係分離自資料記錄區域。 此一格式係不同於圖6中所示之格式。與此情況下，求取 區段未被***資料區域中，且記錄及再生係根據求取結果 (40) 1233104 而被執行。 信號位準波動可隨光學資訊記錄媒體及資訊記錄與再 生裝置之組合而改變。因此，每次一光學資訊記錄媒體被 設定至資訊記錄及再生裝置時（亦即，每次新的光學資訊 被記錄時），則求取區段便透過最內周圍區域及最外周圍 區域中之記錄及再生而被求取。於此情況下，求取區段未 被***資料區域中。因此，記錄及再生係根據求取結果而 被執行，以致其信號位準波動可被恆定地追隨。 接下來，將描述以圖6中所示之格式下的N (= 1 2 8 )區段中之一的求取頻率的求取之結果。亦將描述透過此 一情況下之未知記錄資料的評估結果。每一區段中之多値 資料包含204 8個單元（2048x3位元；八進記錄）。 圖1 1顯示在波形等化之前於個.別多値位準上的再生 信號之分佈。圖1 2顯示在預測性波形等化之後於個別多 値位準上的再生信號之分佈。圖1 3顯示在檢測性波形等 化之後於個別多値位準上的再生信號之分佈。 透過此實施例之波形等化，則個別多値位準上之再生 信號的分佈可被顯著地增進，且再生信號有效地收斂。於 此，預測性等化係數 CO，Cl，C2,及 C3個別爲-0.1 1， 0.41 ’ 0.49,及-0.13。同時，檢測性波形等化之收斂增益G 爲 150。 拾取頭的光學系統具有650 nm之記錄及再生波長；I 、0.65 NA之物鏡、及約0.8微米之光束點直徑BD。單元 長度係0.40微米，而記錄密度爲7.0位元/微米。相較於 (41) 1233104 其可被記錄及再生以相同拾取頭之可重寫光學資訊記錄媒 體（以二元記錄之3.75位元/微米的DVD )，此實施例中 之記錄密度約更高2. 〇倍。以多値信號之最大値及最小値 爲” 1”，則每一多値位準上之分佈的評估係由標準偏差之 平均値σ avg所界定。此實施例之波形等化的效果被相互 比較，使用多値決定中之單元單位的平均値σ avg及誤差 率Err。其比較結果如下： 在波形等化之前： σ avg = 13.5%，Err = 40%至 50% 在預測性波形等化之後： 〇 avg = 3.2%，Err = 8% 至 10% 在檢測性波形等化之後： σ avg = 1.8%，Err = 0% 至 0.2% 上述結果證實其平均値javg減少至1/8而誤差率 E r r減少至1 / 2 0 0或更低，其係由於此實施例之效果。 以此方式，此實施例之多値資料檢測電路預測其位於 待被再生主體記錄標記之前及之後的記錄標記之狀態及影 響程度，當從再生信號再生資訊時，該等再生信號係透過 記錄標記之區域的調變而使其位準被多値化。根據所預測 的影響程度，最佳檢測性波形等化被最終地執行於每一型 態。同時，根據檢測結果，其檢測性波形等化被重複。因 此，碼之間的干擾可透過具有高精確度之波形等化而被消 除0 -45- (42) 1233104 同時，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，等化係數條件被 決定於其包含主體記錄標記前之記錄標記串的已知資料値 、主體記錄標記之已知資料値、及主體記錄標記後之記錄 標記串的已知資料値的三個或更多連續記錄標記之各組合 型態。因此，能夠以高精確度求取其藉由一波形等化操作 以消除碼之間干擾的條件。 再者，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，則包含多値資料 之三個或更多連續項的所有組合之資料串被重複地記錄以 形成多値資料串，且記錄區域被週期性地形成於一分離自 其中記錄有未知多値資料之資料區域的區域中，以致其光 學資訊記錄媒體之信號位準波動可被恆定地追隨。此外， 絕佳的求取效果可被達成，而不管其由於缺陷等之突然改 變。因此，碼之間的干擾可依據信號位準波動而被消除， 且精確的多値決定可被執行。 同時，當欲從再生信號（其係透過記錄標記之區域的 調變而使其位準被多値化）再生資訊時，其指示波形等化 之收斂目標的型態串被記錄爲求取資訊，而求取區域被週 期性地形成於一分離自其中記錄有未知多値資料之資料區 域的區域中，以致其光學資訊記錄媒體之信號位準波動可 被恆定地追隨。此外，絕佳的求取效果可被達成，而不管 其由於缺陷等之突然改變。因此，碼之間的干擾可被消除 ，且精確的多値決定可依據光學資訊記錄媒體之信號位準 -46 - (43) 1233104 變化的頻率特性而被執行。 應注意本發明並不限定於以上明確揭露之實施例，仍 可執行其他更改及修飾而不背離本發明之範圍。 本案係根據對日本專利局於2002年4月1 5日所申請 之日本優先權申請案No. 20〇2-112544，其完整內容被倂 入於此以供參考。 【圖式簡單說明】 圖1顯示一種多値資料檢測電路之結構，其係依據本 發明之多値資料記錄及再生裝置的一實施例； 圖2係顯示於八進（octal )記錄波形等化之情況中的 收斂狀態之一圖形； 圖3係顯示於八進（octal )記錄波形等化之情況中的 收斂狀態之另一圖形； 圖4係顯示依據本發明之一波形等化器的結構之一方 塊圖； 圖5係顯示一種利用圖1所示之多値資料檢測電路的 資訊記錄及再生電路之結構的一方塊圖； 圖6顯示記錄於圖1所示之光學資訊記錄媒體上的記 錄資料之格式； 圖7係圖1所示之多値資料檢測電路中求取預測性等 化係數的操作演算法之一流程圖； 圖8係圖1所示之多値資料檢測電路中求取檢測性等 化係數的操作演算法之一流程圖； -47- (44) 1233104 圖9顯示藉由圖1中所示之多値資料檢測電路的檢測 性波形等化中所使用之等化係數的表列； 圖1 0係再生圖1所示之多値資料檢測電路中的未知 資料之操作演算法的流程圖； 圖1 1顯示在藉由圖1所示之多値資料檢測電路的波 形等化之前於個別多値位準上的再生信號之分佈； 圖1 2顯示在藉由圖1所示之多値資料檢測電路的預 測性波形等化之後於個別多値位準上的再生信號之分佈； 圖1 3顯示在藉由圖1所示之多値資料檢測電路的檢 測性波形等化之後於個別多値位準上的再生信號之分佈； 圖1 4顯示一種多値資料檢測電路之結構，其係本發 明之多値資料記錄及再生裝置的另一實施例； 圖1 5顯示藉由圖1 4中所示之多値資料檢測電路中所 使用之等化係數的表列； 圖1 6係再生圖1 4所示之多値資料檢測電路中的未知 資料之操作演算法的流程圖； 圖1 7係顯示一種習知波形等化器之結構的方塊圖； 圖1 8顯示藉由改變記錄標記相對於稱爲”單元，’之單 位面積的佔有率而實現之多値記錄的一範例； 圖19A及19B顯示一種情況，其中記錄尙未被執行 於其位於主體單元之前及之後的單元中； 圖20A及2〇B顯示一種情況，其中位於主體單元之 前及之後的單元具有與主體單元之記錄標記佔有率相同的 記錄標記佔有率；及 -48- (45) (45)1233104 圖2 1 A及2 1 B顯示一種情況，其中位於主體單元之 前及之後的單元具有最高的記錄標記佔有率。 主要元件對照表 1 :預測器 3 :決定器 1 1 :預測性波形等化器 1 2 :預測性等化係數計算器 1 3 :初始等化係數設定單元 1 4 :預測性臨限値型態檢測器 1 5 :預測性誤差計算器 1 6 :多級資料預測器 1 7 :預測性臨限値產生器 20 :記憶體緩衝器 3 1 :檢測性波形等化器 3 2 :檢測性等化係數及收斂增益計算器 3 3 :檢測性誤差計算器 3 4 :多値資料檢測器 3 5 :檢測性臨限値產生器 3 6 :多値信號記憶體 3 7 :統計處理器 42a-42d :差異計算器 5 〇 :光學資訊記錄媒體 5 1 :拾取頭 -49- (46) (46)1233104 52 : LD驅動器信號產生器 5 3 :多値資料轉變器 5 4 :資訊資料產生器 5 5 :光檢測器 56 : AGC控制器 5 7 :同步信號單元檢測電路 5 8 :取樣信號產生電路 59 :定量AD轉變器 60 :多値信號記憶體 100， 102， 104， 106 ：檢測器 101， 103， 105·•延遲單元 1 1 〇 :預測性波形等化器 1 1 1 :預測性等化係數計算器 1 12 :初始等化係數設定單.元 1 1 3 :預測性臨限値型態檢測器 1 1 4 :預測性誤差計算器 1 1 5 :預測性臨限値產生器 1 1 6 :多値資料預測器 1 1 7 :記憶體緩衝器 1 20 :第一級預測性波形等化器 1 2 1 :檢測性等化係數及收斂增益計算器 122 :檢測性臨限値產生器 123 :檢測性誤差計算器 1 2 4 :統計處理器 -50- (47) 1233104 1 2 5 :多値信號記憶體 126 :第一級多値資料檢測器 127 :記憶體緩衝器 1 3 0 :第二級預測性波形等化器 1 3 1 :檢測性等化係數及收斂增益計算器 1 3 2 :第二級多値資料檢測器 1 3 3 :記憶體緩衝器 1 40 :第三級預測性波形等化器 1 4 1 :檢測性等化係數及收斂增益計算器 1 42 :第三級多値資料檢測器(al + a2 + a3 + a4 + a5 + a6 + a7 + a8) / 8 It averages each frame regardless of the known frame recorded by the subject to be reproduced. As a result, the list of equalization coefficients used in this embodiment becomes a list containing 6 4x4 = 2 56 coefficients. The obtained equalization coefficients are the same in all multi-determination circuits configured in series. Next, an example of the reproduction operation of unknown data performed by this multiple data detection circuit will be described. In the multiple data detection circuit of this embodiment, the detection waveform equalization and determination need to be repeated three times before the determination result of the multiple data is converged. Figure 16 is a flowchart of an operation algorithm using this multi-data detection circuit to reproduce one of the unknown data. Based on the results of the predictive equalization coefficients and the results of the detective equalization coefficients, it is decided to be executed on an unknown data string. Here, a lot of data in the seeking section in the format shown in FIG. 6 are used as the known data -42- (39) 1233104 data. In the format shown in FIG. 6, a seek section is inserted into every N consecutive unknown data sections. Therefore, according to the result of retrieving the segment, the number of unknown data segments is decided to be executed. The insertion frequency of the determination section (one determination section is inserted into each of the N unknown data sections) is determined by the frequency characteristics of the signal level fluctuation of the optical information recording medium. If the signal level fluctuation is generated every week of the optical information recording medium 50, for example, at least four determination sections need to be inserted within one week to follow the fluctuation frequency. If a signal level fluctuation is generated in each radial area, a seeking section should be inserted within every few micrometers of each radial area to follow the fluctuation. When it is desired to update the results, it is better to add the previous results and newly reproduced information to the results, so that the results will not be adversely affected by sudden changes in defects. In a case where the calculation result is calculated through a statistical operation on 16 calculation sections, for example, "the information of the first calculation section of the 16 calculation sections" is replaced With "recently reproduced seeking section information, 'in statistical operations. In this way, the latest finding results can be reflected in the updated finding results, and from" due to the sudden change of the defect, the disadvantage Impact can be avoided in the evaluation procedure. If the signal level fluctuation is generated every time a new optical information recording medium 50 is set, the obtained section should be recorded and only reproduced from the innermost peripheral area and the outermost peripheral area, which is separated from Data recording area. This format is different from the format shown in FIG. In this case, the seeking section is not inserted into the data area, and recording and reproduction are performed based on the finding result (40) 1233104. The signal level fluctuation can be changed by the optical information recording medium and the combination of the information recording and reproduction device. Therefore, each time an optical information recording medium is set to the information recording and reproducing device (that is, each time new optical information is recorded), the seeking section passes through the innermost peripheral area and the outermost peripheral area. Recording and reproduction are required. In this case, the seek section is not inserted into the data area. Therefore, the recording and reproduction are performed based on the obtained result, so that the fluctuation of its signal level can be constantly followed. Next, the result of the determination of the determination frequency in one of the N (= 1 2 8) sections in the format shown in FIG. 6 will be described. The results of the assessment through unknown recorded data in this case will also be described. The data in each section contains 204 8 units (2048x3 bits; octal records). Figure 11 shows the distribution of the reproduced signal at the different levels before the waveform is equalized. Figure 12 shows the distribution of the reproduced signals at the individual multi-levels after the prediction waveform is equalized. Figure 13 shows the distribution of the reproduced signal at the individual multi-levels after the detection waveform is equalized. By equalizing the waveforms of this embodiment, the distribution of the reproduced signals at individual multi-levels can be significantly improved, and the reproduced signals effectively converge. Here, the predictive equalization coefficients CO, Cl, C2, and C3 are individually -0.1 1, 0.41 '0.49, and -0.13. At the same time, the convergence gain G of the detection waveform equalization is 150. The optical system of the pickup head has a recording and reproduction wavelength of 650 nm; I, an objective lens of 0.65 NA, and a beam spot diameter BD of about 0.8 microns. The cell length is 0.40 microns and the recording density is 7.0 bits / micron. Compared to (41) 1233104, which can be recorded and reproduced with a rewritable optical information recording medium with the same pickup head (a 3.75 bit / micron DVD recorded in binary), the recording density in this embodiment is about higher 2. 〇 times. Taking the maximum and minimum values of the multi-signal as "1", the evaluation of the distribution at each multi-level is defined by the mean 値 σ avg of the standard deviation. The effects of the equalization of the waveforms in this embodiment are compared with each other, and the average 値 σ avg and the error rate Err of the unit units in the multiple determination are used. The comparison results are as follows: Before waveform equalization: σ avg = 13.5%, Err = 40% to 50% After predictive waveform equalization: 〇avg = 3.2%, Err = 8% to 10% In detection waveform, etc. After conversion: σ avg = 1.8%, Err = 0% to 0.2% The above results confirm that its average 値 javg is reduced to 1/8 and the error rate Err is reduced to 1/20 0 or lower, which is due to this embodiment The effect. In this way, the multiple data detection circuit of this embodiment predicts the state and influence of the recording marks before and after the recording marks of the subject to be reproduced. When information is reproduced from the reproduced signals, the reproduced signals are transmitted through the recorded marks The adjustment of the area makes its level multiplied. Based on the predicted degree of influence, the best detectable waveform equalization is ultimately performed for each type. At the same time, the detection waveform equalization is repeated based on the detection results. Therefore, the interference between codes can be eliminated by equalizing the waveform with high accuracy. 0 -45- (42) 1233104 At the same time, when the signal When the information is reproduced, the equalization coefficient condition is determined by including the known data of the record mark string before the main record mark, the known data of the main record mark, and the record mark after the main record mark. Each combination of three or more consecutive record marks of a string of known data. Therefore, it is possible to obtain a condition with high accuracy to eliminate interference between codes by a waveform equalization operation. Furthermore, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recorded mark), it contains all combinations of three or more consecutive items of multi-data The data string is repeatedly recorded to form multiple data strings, and the recording area is periodically formed in an area separated from the data area in which unknown multiple data is recorded, so that the signal level of its optical information recording medium fluctuates. Can be constantly followed. In addition, excellent results can be achieved regardless of sudden changes due to defects or the like. As a result, interference between codes can be eliminated based on signal level fluctuations, and accurate multiple decision can be performed. At the same time, when information is to be reproduced from a reproduction signal (which is multiplied by the modulation of the area of the recording mark), the pattern string indicating the convergence target of the waveform equalization is recorded as the information seeking The seeking area is periodically formed in an area separated from the data area in which unknown data is recorded, so that the signal level fluctuations of its optical information recording medium can be constantly followed. In addition, excellent results can be achieved regardless of sudden changes due to defects or the like. Therefore, the interference between the codes can be eliminated, and the precise multi-decision can be performed according to the frequency characteristics of the signal level of the optical information recording medium -46-(43) 1233104. It should be noted that the present invention is not limited to the embodiments explicitly disclosed above, and other changes and modifications can be performed without departing from the scope of the present invention. This case is based on Japanese Priority Application No. 200-112544 filed by the Japan Patent Office on April 15, 2002, the entire contents of which are incorporated herein by reference. [Brief description of the figure] FIG. 1 shows the structure of a multi-channel data detection circuit, which is an embodiment of a multi-channel data recording and reproduction device according to the present invention; FIG. 2 shows an octal recording waveform equalization Fig. 3 is another graph showing the state of convergence in the case of octal recording waveform equalization; Fig. 4 is a diagram showing the structure of a waveform equalizer according to the present invention A block diagram; FIG. 5 is a block diagram showing the structure of an information recording and reproduction circuit using the multi-data detection circuit shown in FIG. 1; FIG. 6 shows a recording on the optical information recording medium shown in FIG. Format of recorded data; FIG. 7 is a flowchart of an operation algorithm for obtaining predictive equalization coefficients in the multi-data detection circuit shown in FIG. 1; FIG. 8 is a flowchart of the multi-data detection circuit shown in FIG. One flow chart of the operational algorithm for obtaining the detection equalization coefficient; -47- (44) 1233104 FIG. 9 shows the equalization used in the detection waveform equalization by the multiple data detection circuit shown in FIG. 1 Table of coefficients; Figure 1 0 is a flowchart of an operation algorithm for reproducing unknown data in the multi-data detection circuit shown in FIG. 1; FIG. 11 shows that prior to waveform equalization by the multi-data detection circuit shown in FIG. The distribution of the reproduced signal at the level; Figure 12 shows the distribution of the reproduced signal at the individual multiple levels after equalization by the predictive waveform of the multiple data detection circuit shown in Figure 1; Figure 1 3 Shows the distribution of the reproduced signals at the individual multi-levels after being equalized by the detection waveform of the multi-level data detection circuit shown in Figure 1. Figure 14 shows the structure of a multi-level data detection circuit, which is a Another embodiment of the multi-data recording and reproducing device of the invention; FIG. 15 shows a table of equalization coefficients used in the multi-data detection circuit shown in FIG. 14; FIG. 16 is a reproduction chart Figure 14 shows the flowchart of the operation algorithm of unknown data in the multi-data detection circuit. Figure 17 is a block diagram showing the structure of a conventional waveform equalizer. Figure 18 shows the relative change by changing the record mark. Unit area An example of multiple records achieved by occupancy; Figures 19A and 19B show a situation where recording is not performed in units that are located before and after the main unit; Figures 20A and 20B show a situation where The units before and after the main unit have the same record mark occupancy as the record mark occupancy of the main unit; and -48- (45) (45) 1233104 Figure 2 1 A and 2 1 B show a situation where the main unit is located The units before and after have the highest record mark occupancy rate. Table 1 Comparison of main components: Predictor 3: Determinator 1 1: Predictive waveform equalizer 1 2: Predictive equalization coefficient calculator 1 3: Initial equalization Coefficient setting unit 14: predictive threshold threshold detector 15: predictive error calculator 16: multi-level data predictor 17: predictive threshold threshold generator 20: memory buffer 3 1: Detector waveform equalizer 3 2: Detector equalization coefficient and convergence gain calculator 3 3: Detector error calculator 3 4: Multiple data detector 3 5: Detectable threshold generator 3 6: Multiple Signal memory 37: statistical processing 42a-42d: Difference calculator 5 〇: Optical information recording medium 5 1: Pickup head -49- (46) (46) 1233104 52: LD drive signal generator 5 3: Multi-data converter 5 4: Information data generation 5: Photodetector 56: AGC controller 5 7: Synchronous signal unit detection circuit 5 8: Sampling signal generation circuit 59: Quantitative AD converter 60: Multi-signal memory 100, 102, 104, 106: Detector 101, 103, 105 · • Delay unit 1 1 0: Predictive waveform equalizer 1 1 1: Predictive equalization coefficient calculator 1 12: Initial equalization coefficient setting sheet. Element 1 1 3: Predictive threshold 値Type detector 1 1 4: Predictive error calculator 1 1 5: Predictive threshold threshold generator 1 1 6: Multi-data predictor 1 1 7: Memory buffer 1 20: First-level predictive waveform Equalizer 1 2 1: Detectable equalization coefficient and convergence gain calculator 122: Detectable threshold generator 123: Detectable error calculator 1 2 4: Statistical processor-50- (47) 1233104 1 2 5 : Multi-channel signal memory 126: First-level multi-channel data detector 127: Memory buffer 1 3 0: Second-level predictive waveform, etc. 1 3 1: Detective equalization coefficient and convergence gain calculator 1 3 2: Second-level multi-data detector 1 3 3: Memory buffer 1 40: Third-level predictive waveform equalizer 1 4 1 : Detective Equalization Coefficient and Convergence Gain Calculator 1 42: Third-level multi-data detector
-51 --51-